chore(deps): bump github.com/swaggo/echo-swagger from 1.4.1 to 1.5.0

Bumps [github.com/swaggo/echo-swagger](https://github.com/swaggo/echo-swagger) from 1.4.1 to 1.5.0.
- [Release notes](https://github.com/swaggo/echo-swagger/releases)
- [Commits](https://github.com/swaggo/echo-swagger/compare/v1.4.1...v1.5.0)

---
updated-dependencies:
- dependency-name: github.com/swaggo/echo-swagger
  dependency-version: 1.5.0
  dependency-type: direct:production
  update-type: version-update:semver-minor
...

Signed-off-by: dependabot[bot] <support@github.com>

.github/workflows/bump_deps.yaml

@@ -18,10 +18,6 @@ jobs:
             variable: "WHISPER_CPP_VERSION"
             branch: "master"
             file: "backend/go/whisper/Makefile"
-          - repository: "PABannier/bark.cpp"
-            variable: "BARKCPP_VERSION"
-            branch: "main"
-            file: "Makefile"
           - repository: "leejet/stable-diffusion.cpp"
             variable: "STABLEDIFFUSION_GGML_VERSION"
             branch: "master"

.github/workflows/release.yaml

@@ -18,7 +18,7 @@ jobs:
         with:
           go-version: 1.23
       - name: Run GoReleaser
-        uses: goreleaser/goreleaser-action@v6
+        uses: goreleaser/goreleaser-action@v7
         with:
           version: v2.11.0
           args: release --clean

.github/workflows/stalebot.yml

@@ -11,7 +11,7 @@ jobs:
     if: github.repository == 'mudler/LocalAI'
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/stale@997185467fa4f803885201cee163a9f38240193d # v9
+      - uses: actions/stale@b5d41d4e1d5dceea10e7104786b73624c18a190f # v9
         with:
           stale-issue-message: 'This issue is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 5 days.'
           stale-pr-message: 'This PR is stale because it has been open 90 days with no activity. Remove stale label or comment or this will be closed in 10 days.'

CLAUDE.md

@@ -0,0 +1 @@
+AGENTS.md

backend/cpp/llama-cpp/Makefile

@@ -1,5 +1,5 @@
 
-LLAMA_VERSION?=11c325c6e0666a30590cde390d5746a405e536b9
+LLAMA_VERSION?=b68a83e641b3ebe6465970b34e99f3f0e0a0b21a
 LLAMA_REPO?=https://github.com/ggerganov/llama.cpp
 
 CMAKE_ARGS?=

core/backend/embeddings.go

@@ -2,8 +2,10 @@ package backend
 
 import (
 	"fmt"
+	"time"
 
 	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/trace"
 
 	"github.com/mudler/LocalAI/pkg/grpc"
 	model "github.com/mudler/LocalAI/pkg/model"
@@ -53,7 +55,7 @@ func ModelEmbedding(s string, tokens []int, loader *model.ModelLoader, modelConf
 		}
 	}
 
-	return func() ([]float32, error) {
+	wrappedFn := func() ([]float32, error) {
 		embeds, err := fn()
 		if err != nil {
 			return embeds, err
@@ -67,5 +69,48 @@ func ModelEmbedding(s string, tokens []int, loader *model.ModelLoader, modelConf
 			}
 		}
 		return embeds, nil
-	}, nil
+	}
+
+	if appConfig.EnableTracing {
+		trace.InitBackendTracingIfEnabled(appConfig.TracingMaxItems)
+
+		traceData := map[string]any{
+			"input_text":         trace.TruncateString(s, 1000),
+			"input_tokens_count": len(tokens),
+		}
+
+		startTime := time.Now()
+		originalFn := wrappedFn
+		wrappedFn = func() ([]float32, error) {
+			result, err := originalFn()
+			duration := time.Since(startTime)
+
+			traceData["embedding_dimensions"] = len(result)
+
+			errStr := ""
+			if err != nil {
+				errStr = err.Error()
+			}
+
+			summary := trace.TruncateString(s, 200)
+			if summary == "" {
+				summary = fmt.Sprintf("tokens[%d]", len(tokens))
+			}
+
+			trace.RecordBackendTrace(trace.BackendTrace{
+				Timestamp: startTime,
+				Duration:  duration,
+				Type:      trace.BackendTraceEmbedding,
+				ModelName: modelConfig.Name,
+				Backend:   modelConfig.Backend,
+				Summary:   summary,
+				Error:     errStr,
+				Data:      traceData,
+			})
+
+			return result, err
+		}
+	}
+
+	return wrappedFn, nil
 }

core/backend/image.go

@@ -1,7 +1,10 @@
 package backend
 
 import (
+	"time"
+
 	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/trace"
 
 	"github.com/mudler/LocalAI/pkg/grpc/proto"
 	model "github.com/mudler/LocalAI/pkg/model"
@@ -36,6 +39,46 @@ func ImageGeneration(height, width, step, seed int, positive_prompt, negative_pr
 		return err
 	}
 
+	if appConfig.EnableTracing {
+		trace.InitBackendTracingIfEnabled(appConfig.TracingMaxItems)
+
+		traceData := map[string]any{
+			"positive_prompt": positive_prompt,
+			"negative_prompt": negative_prompt,
+			"height":          height,
+			"width":           width,
+			"step":            step,
+			"seed":            seed,
+			"source_image":    src,
+			"destination":     dst,
+		}
+
+		startTime := time.Now()
+		originalFn := fn
+		fn = func() error {
+			err := originalFn()
+			duration := time.Since(startTime)
+
+			errStr := ""
+			if err != nil {
+				errStr = err.Error()
+			}
+
+			trace.RecordBackendTrace(trace.BackendTrace{
+				Timestamp: startTime,
+				Duration:  duration,
+				Type:      trace.BackendTraceImageGeneration,
+				ModelName: modelConfig.Name,
+				Backend:   modelConfig.Backend,
+				Summary:   trace.TruncateString(positive_prompt, 200),
+				Error:     errStr,
+				Data:      traceData,
+			})
+
+			return err
+		}
+	}
+
 	return fn, nil
 }
 

core/backend/llm.go

@@ -7,11 +7,13 @@ import (
 	"slices"
 	"strings"
 	"sync"
+	"time"
 	"unicode/utf8"
 
 	"github.com/mudler/xlog"
 
 	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/trace"
 	"github.com/mudler/LocalAI/core/schema"
 	"github.com/mudler/LocalAI/core/services"
 
@@ -220,6 +222,84 @@ func ModelInference(ctx context.Context, s string, messages schema.Messages, ima
 		}
 	}
 
+	if o.EnableTracing {
+		trace.InitBackendTracingIfEnabled(o.TracingMaxItems)
+
+		traceData := map[string]any{
+			"prompt":                 s,
+			"use_tokenizer_template": c.TemplateConfig.UseTokenizerTemplate,
+			"chat_template":          c.TemplateConfig.Chat,
+			"function_template":      c.TemplateConfig.Functions,
+			"grammar":               c.Grammar,
+			"stop_words":            c.StopWords,
+			"streaming":             tokenCallback != nil,
+			"images_count":          len(images),
+			"videos_count":          len(videos),
+			"audios_count":          len(audios),
+		}
+
+		if len(messages) > 0 {
+			if msgJSON, err := json.Marshal(messages); err == nil {
+				traceData["messages"] = string(msgJSON)
+			}
+		}
+		if tools != "" {
+			traceData["tools"] = tools
+		}
+		if toolChoice != "" {
+			traceData["tool_choice"] = toolChoice
+		}
+		if reasoningJSON, err := json.Marshal(c.ReasoningConfig); err == nil {
+			traceData["reasoning_config"] = string(reasoningJSON)
+		}
+		traceData["functions_config"] = map[string]any{
+			"grammar_disabled":  c.FunctionsConfig.GrammarConfig.NoGrammar,
+			"parallel_calls":    c.FunctionsConfig.GrammarConfig.ParallelCalls,
+			"mixed_mode":        c.FunctionsConfig.GrammarConfig.MixedMode,
+			"xml_format_preset": c.FunctionsConfig.XMLFormatPreset,
+		}
+		if c.Temperature != nil {
+			traceData["temperature"] = *c.Temperature
+		}
+		if c.TopP != nil {
+			traceData["top_p"] = *c.TopP
+		}
+		if c.Maxtokens != nil {
+			traceData["max_tokens"] = *c.Maxtokens
+		}
+
+		startTime := time.Now()
+		originalFn := fn
+		fn = func() (LLMResponse, error) {
+			resp, err := originalFn()
+			duration := time.Since(startTime)
+
+			traceData["response"] = resp.Response
+			traceData["token_usage"] = map[string]any{
+				"prompt":     resp.Usage.Prompt,
+				"completion": resp.Usage.Completion,
+			}
+
+			errStr := ""
+			if err != nil {
+				errStr = err.Error()
+			}
+
+			trace.RecordBackendTrace(trace.BackendTrace{
+				Timestamp: startTime,
+				Duration:  duration,
+				Type:      trace.BackendTraceLLM,
+				ModelName: c.Name,
+				Backend:   c.Backend,
+				Summary:   trace.GenerateLLMSummary(messages, s),
+				Error:     errStr,
+				Data:      traceData,
+			})
+
+			return resp, err
+		}
+	}
+
 	return fn, nil
 }
 

core/backend/rerank.go

@@ -3,8 +3,10 @@ package backend
 import (
 	"context"
 	"fmt"
+	"time"
 
 	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/trace"
 	"github.com/mudler/LocalAI/pkg/grpc/proto"
 	model "github.com/mudler/LocalAI/pkg/model"
 )
@@ -20,7 +22,35 @@ func Rerank(request *proto.RerankRequest, loader *model.ModelLoader, appConfig *
 		return nil, fmt.Errorf("could not load rerank model")
 	}
 
+	var startTime time.Time
+	if appConfig.EnableTracing {
+		trace.InitBackendTracingIfEnabled(appConfig.TracingMaxItems)
+		startTime = time.Now()
+	}
+
 	res, err := rerankModel.Rerank(context.Background(), request)
 
+	if appConfig.EnableTracing {
+		errStr := ""
+		if err != nil {
+			errStr = err.Error()
+		}
+
+		trace.RecordBackendTrace(trace.BackendTrace{
+			Timestamp: startTime,
+			Duration:  time.Since(startTime),
+			Type:      trace.BackendTraceRerank,
+			ModelName: modelConfig.Name,
+			Backend:   modelConfig.Backend,
+			Summary:   trace.TruncateString(request.Query, 200),
+			Error:     errStr,
+			Data: map[string]any{
+				"query":           request.Query,
+				"documents_count": len(request.Documents),
+				"top_n":           request.TopN,
+			},
+		})
+	}
+
 	return res, err
 }

core/backend/soundgeneration.go

@@ -5,8 +5,10 @@ import (
 	"fmt"
 	"os"
 	"path/filepath"
+	"time"
 
 	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/trace"
 	"github.com/mudler/LocalAI/pkg/grpc/proto"
 	"github.com/mudler/LocalAI/pkg/model"
 	"github.com/mudler/LocalAI/pkg/utils"
@@ -92,7 +94,51 @@ func SoundGeneration(
 		req.Instrumental = instrumental
 	}
 
+	var startTime time.Time
+	if appConfig.EnableTracing {
+		trace.InitBackendTracingIfEnabled(appConfig.TracingMaxItems)
+		startTime = time.Now()
+	}
+
 	res, err := soundGenModel.SoundGeneration(context.Background(), req)
+
+	if appConfig.EnableTracing {
+		errStr := ""
+		if err != nil {
+			errStr = err.Error()
+		} else if res != nil && !res.Success {
+			errStr = fmt.Sprintf("sound generation error: %s", res.Message)
+		}
+
+		summary := trace.TruncateString(text, 200)
+		if summary == "" && caption != "" {
+			summary = trace.TruncateString(caption, 200)
+		}
+
+		traceData := map[string]any{
+			"text":    text,
+			"caption": caption,
+			"lyrics":  lyrics,
+		}
+		if duration != nil {
+			traceData["duration"] = *duration
+		}
+		if temperature != nil {
+			traceData["temperature"] = *temperature
+		}
+
+		trace.RecordBackendTrace(trace.BackendTrace{
+			Timestamp: startTime,
+			Duration:  time.Since(startTime),
+			Type:      trace.BackendTraceSoundGeneration,
+			ModelName: modelConfig.Name,
+			Backend:   modelConfig.Backend,
+			Summary:   summary,
+			Error:     errStr,
+			Data:      traceData,
+		})
+	}
+
 	if err != nil {
 		return "", nil, err
 	}

core/backend/tokenize.go

@@ -1,7 +1,10 @@
 package backend
 
 import (
+	"time"
+
 	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/trace"
 	"github.com/mudler/LocalAI/core/schema"
 	"github.com/mudler/LocalAI/pkg/grpc"
 	"github.com/mudler/LocalAI/pkg/model"
@@ -21,8 +24,41 @@ func ModelTokenize(s string, loader *model.ModelLoader, modelConfig config.Model
 	predictOptions := gRPCPredictOpts(modelConfig, loader.ModelPath)
 	predictOptions.Prompt = s
 
+	var startTime time.Time
+	if appConfig.EnableTracing {
+		trace.InitBackendTracingIfEnabled(appConfig.TracingMaxItems)
+		startTime = time.Now()
+	}
+
 	// tokenize the string
 	resp, err := inferenceModel.TokenizeString(appConfig.Context, predictOptions)
+
+	if appConfig.EnableTracing {
+		errStr := ""
+		if err != nil {
+			errStr = err.Error()
+		}
+
+		tokenCount := 0
+		if resp.Tokens != nil {
+			tokenCount = len(resp.Tokens)
+		}
+
+		trace.RecordBackendTrace(trace.BackendTrace{
+			Timestamp: startTime,
+			Duration:  time.Since(startTime),
+			Type:      trace.BackendTraceTokenize,
+			ModelName: modelConfig.Name,
+			Backend:   modelConfig.Backend,
+			Summary:   trace.TruncateString(s, 200),
+			Error:     errStr,
+			Data: map[string]any{
+				"input_text":  trace.TruncateString(s, 1000),
+				"token_count": tokenCount,
+			},
+		})
+	}
+
 	if err != nil {
 		return schema.TokenizeResponse{}, err
 	}

core/backend/transcript.go

@@ -6,6 +6,7 @@ import (
 	"time"
 
 	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/trace"
 	"github.com/mudler/LocalAI/core/schema"
 
 	"github.com/mudler/LocalAI/pkg/grpc/proto"
@@ -28,6 +29,12 @@ func ModelTranscription(audio, language string, translate, diarize bool, prompt
 		return nil, fmt.Errorf("could not load transcription model")
 	}
 
+	var startTime time.Time
+	if appConfig.EnableTracing {
+		trace.InitBackendTracingIfEnabled(appConfig.TracingMaxItems)
+		startTime = time.Now()
+	}
+
 	r, err := transcriptionModel.AudioTranscription(context.Background(), &proto.TranscriptRequest{
 		Dst:       audio,
 		Language:  language,
@@ -37,6 +44,24 @@ func ModelTranscription(audio, language string, translate, diarize bool, prompt
 		Prompt:    prompt,
 	})
 	if err != nil {
+		if appConfig.EnableTracing {
+			trace.RecordBackendTrace(trace.BackendTrace{
+				Timestamp: startTime,
+				Duration:  time.Since(startTime),
+				Type:      trace.BackendTraceTranscription,
+				ModelName: modelConfig.Name,
+				Backend:   modelConfig.Backend,
+				Summary:   trace.TruncateString(audio, 200),
+				Error:     err.Error(),
+				Data: map[string]any{
+					"audio_file": audio,
+					"language":   language,
+					"translate":  translate,
+					"diarize":    diarize,
+					"prompt":     prompt,
+				},
+			})
+		}
 		return nil, err
 	}
 	tr := &schema.TranscriptionResult{
@@ -57,5 +82,26 @@ func ModelTranscription(audio, language string, translate, diarize bool, prompt
 				Speaker: s.Speaker,
 			})
 	}
+
+	if appConfig.EnableTracing {
+		trace.RecordBackendTrace(trace.BackendTrace{
+			Timestamp: startTime,
+			Duration:  time.Since(startTime),
+			Type:      trace.BackendTraceTranscription,
+			ModelName: modelConfig.Name,
+			Backend:   modelConfig.Backend,
+			Summary:   trace.TruncateString(audio+" -> "+tr.Text, 200),
+			Data: map[string]any{
+				"audio_file":     audio,
+				"language":       language,
+				"translate":      translate,
+				"diarize":        diarize,
+				"prompt":         prompt,
+				"result_text":    tr.Text,
+				"segments_count": len(tr.Segments),
+			},
+		})
+	}
+
 	return tr, err
 }

core/backend/tts.go

@@ -8,8 +8,10 @@ import (
 	"fmt"
 	"os"
 	"path/filepath"
+	"time"
 
 	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/trace"
 	laudio "github.com/mudler/LocalAI/pkg/audio"
 
 	"github.com/mudler/LocalAI/pkg/grpc/proto"
@@ -60,6 +62,12 @@ func ModelTTS(
 		modelPath = modelConfig.Model // skip this step if it fails?????
 	}
 
+	var startTime time.Time
+	if appConfig.EnableTracing {
+		trace.InitBackendTracingIfEnabled(appConfig.TracingMaxItems)
+		startTime = time.Now()
+	}
+
 	res, err := ttsModel.TTS(context.Background(), &proto.TTSRequest{
 		Text:     text,
 		Model:    modelPath,
@@ -67,6 +75,31 @@ func ModelTTS(
 		Dst:      filePath,
 		Language: &language,
 	})
+
+	if appConfig.EnableTracing {
+		errStr := ""
+		if err != nil {
+			errStr = err.Error()
+		} else if !res.Success {
+			errStr = fmt.Sprintf("TTS error: %s", res.Message)
+		}
+
+		trace.RecordBackendTrace(trace.BackendTrace{
+			Timestamp: startTime,
+			Duration:  time.Since(startTime),
+			Type:      trace.BackendTraceTTS,
+			ModelName: modelConfig.Name,
+			Backend:   modelConfig.Backend,
+			Summary:   trace.TruncateString(text, 200),
+			Error:     errStr,
+			Data: map[string]any{
+				"text":     text,
+				"voice":    voice,
+				"language": language,
+			},
+		})
+	}
+
 	if err != nil {
 		return "", nil, err
 	}
@@ -115,6 +148,12 @@ func ModelTTSStream(
 		modelPath = modelConfig.Model // skip this step if it fails?????
 	}
 
+	var startTime time.Time
+	if appConfig.EnableTracing {
+		trace.InitBackendTracingIfEnabled(appConfig.TracingMaxItems)
+		startTime = time.Now()
+	}
+
 	var sampleRate uint32 = 16000 // default
 	headerSent := false
 	var callbackErr error
@@ -171,6 +210,34 @@ func ModelTTSStream(
 		}
 	})
 
+	resultErr := err
+	if callbackErr != nil {
+		resultErr = callbackErr
+	}
+
+	if appConfig.EnableTracing {
+		errStr := ""
+		if resultErr != nil {
+			errStr = resultErr.Error()
+		}
+
+		trace.RecordBackendTrace(trace.BackendTrace{
+			Timestamp: startTime,
+			Duration:  time.Since(startTime),
+			Type:      trace.BackendTraceTTS,
+			ModelName: modelConfig.Name,
+			Backend:   modelConfig.Backend,
+			Summary:   trace.TruncateString(text, 200),
+			Error:     errStr,
+			Data: map[string]any{
+				"text":      text,
+				"voice":     voice,
+				"language":  language,
+				"streaming": true,
+			},
+		})
+	}
+
 	if callbackErr != nil {
 		return callbackErr
 	}

core/backend/video.go

@@ -1,7 +1,10 @@
 package backend
 
 import (
+	"time"
+
 	"github.com/mudler/LocalAI/core/config"
+	"github.com/mudler/LocalAI/core/trace"
 
 	"github.com/mudler/LocalAI/pkg/grpc/proto"
 	model "github.com/mudler/LocalAI/pkg/model"
@@ -37,5 +40,46 @@ func VideoGeneration(height, width int32, prompt, negativePrompt, startImage, en
 		return err
 	}
 
+	if appConfig.EnableTracing {
+		trace.InitBackendTracingIfEnabled(appConfig.TracingMaxItems)
+
+		traceData := map[string]any{
+			"prompt":          prompt,
+			"negative_prompt": negativePrompt,
+			"height":          height,
+			"width":           width,
+			"num_frames":      numFrames,
+			"fps":             fps,
+			"seed":            seed,
+			"cfg_scale":       cfgScale,
+			"step":            step,
+		}
+
+		startTime := time.Now()
+		originalFn := fn
+		fn = func() error {
+			err := originalFn()
+			duration := time.Since(startTime)
+
+			errStr := ""
+			if err != nil {
+				errStr = err.Error()
+			}
+
+			trace.RecordBackendTrace(trace.BackendTrace{
+				Timestamp: startTime,
+				Duration:  duration,
+				Type:      trace.BackendTraceVideoGeneration,
+				ModelName: modelConfig.Name,
+				Backend:   modelConfig.Backend,
+				Summary:   trace.TruncateString(prompt, 200),
+				Error:     errStr,
+				Data:      traceData,
+			})
+
+			return err
+		}
+	}
+
 	return fn, nil
 }

core/cli/run.go

@@ -71,6 +71,7 @@ type RunCMD struct {
 	WatchdogIdleTimeout                string   `env:"LOCALAI_WATCHDOG_IDLE_TIMEOUT,WATCHDOG_IDLE_TIMEOUT" default:"15m" help:"Threshold beyond which an idle backend should be stopped" group:"backends"`
 	EnableWatchdogBusy                 bool     `env:"LOCALAI_WATCHDOG_BUSY,WATCHDOG_BUSY" default:"false" help:"Enable watchdog for stopping backends that are busy longer than the watchdog-busy-timeout" group:"backends"`
 	WatchdogBusyTimeout                string   `env:"LOCALAI_WATCHDOG_BUSY_TIMEOUT,WATCHDOG_BUSY_TIMEOUT" default:"5m" help:"Threshold beyond which a busy backend should be stopped" group:"backends"`
+	WatchdogInterval                   string   `env:"LOCALAI_WATCHDOG_INTERVAL,WATCHDOG_INTERVAL" default:"500ms" help:"Interval between watchdog checks (e.g., 500ms, 5s, 1m) (default: 500ms)" group:"backends"`
 	EnableMemoryReclaimer              bool     `env:"LOCALAI_MEMORY_RECLAIMER,MEMORY_RECLAIMER,LOCALAI_GPU_RECLAIMER,GPU_RECLAIMER" default:"false" help:"Enable memory threshold monitoring to auto-evict backends when memory usage exceeds threshold (uses GPU VRAM if available, otherwise RAM)" group:"backends"`
 	MemoryReclaimerThreshold           float64  `env:"LOCALAI_MEMORY_RECLAIMER_THRESHOLD,MEMORY_RECLAIMER_THRESHOLD,LOCALAI_GPU_RECLAIMER_THRESHOLD,GPU_RECLAIMER_THRESHOLD" default:"0.95" help:"Memory usage threshold (0.0-1.0) that triggers backend eviction (default 0.95 = 95%%)" group:"backends"`
 	ForceEvictionWhenBusy              bool     `env:"LOCALAI_FORCE_EVICTION_WHEN_BUSY,FORCE_EVICTION_WHEN_BUSY" default:"false" help:"Force eviction even when models have active API calls (default: false for safety)" group:"backends"`
@@ -215,6 +216,13 @@ func (r *RunCMD) Run(ctx *cliContext.Context) error {
 			}
 			opts = append(opts, config.SetWatchDogBusyTimeout(dur))
 		}
+		if r.WatchdogInterval != "" {
+			dur, err := time.ParseDuration(r.WatchdogInterval)
+			if err != nil {
+				return err
+			}
+			opts = append(opts, config.SetWatchDogInterval(dur))
+		}
 	}
 
 	// Handle memory reclaimer (uses GPU VRAM if available, otherwise RAM)

core/cli/transcript.go

@@ -31,8 +31,8 @@ type TranscriptCMD struct {
 	ModelsPath       string                                 `env:"LOCALAI_MODELS_PATH,MODELS_PATH" type:"path" default:"${basepath}/models" help:"Path containing models used for inferencing" group:"storage"`
 	BackendGalleries string                                 `env:"LOCALAI_BACKEND_GALLERIES,BACKEND_GALLERIES" help:"JSON list of backend galleries" group:"backends" default:"${backends}"`
 	Prompt           string                                 `short:"p" help:"Previous transcribed text or words that hint at what the model should expect"`
-	ResponseFormat   schema.TranscriptionResponseFormatType `short:"f" default:"" help:"Response format for Whisper models, can be one of (txt, lrc, srt, vtt, json, json_verbose)"`
-	PrettyPrint      bool                                   `help:"Used with response_format json or json_verbose for pretty printing"`
+	ResponseFormat   schema.TranscriptionResponseFormatType `short:"f" default:"" help:"Response format for Whisper models, can be one of (txt, lrc, srt, vtt, json, verbose_json)"`
+	PrettyPrint      bool                                   `help:"Used with response_format json or verbose_json for pretty printing"`
 }
 
 func (t *TranscriptCMD) Run(ctx *cliContext.Context) error {

core/config/application_config.go

@@ -98,10 +98,11 @@ func NewApplicationConfig(o ...AppOption) *ApplicationConfig {
 		Context:                  context.Background(),
 		UploadLimitMB:            15,
 		Debug:                    true,
-		AgentJobRetentionDays:    30,              // Default: 30 days
-		LRUEvictionMaxRetries:    30,              // Default: 30 retries
-		LRUEvictionRetryInterval: 1 * time.Second, // Default: 1 second
-		TracingMaxItems:       1024,
+		AgentJobRetentionDays:    30,                     // Default: 30 days
+		LRUEvictionMaxRetries:    30,                     // Default: 30 retries
+		LRUEvictionRetryInterval: 1 * time.Second,        // Default: 1 second
+		WatchDogInterval:         500 * time.Millisecond, // Default: 500ms
+		TracingMaxItems:          1024,
 		PathWithoutAuth: []string{
 			"/static/",
 			"/generated-audio/",
@@ -208,6 +209,12 @@ func SetWatchDogIdleTimeout(t time.Duration) AppOption {
 	}
 }
 
+func SetWatchDogInterval(t time.Duration) AppOption {
+	return func(o *ApplicationConfig) {
+		o.WatchDogInterval = t
+	}
+}
+
 // EnableMemoryReclaimer enables memory threshold monitoring.
 // When enabled, the watchdog will evict backends if memory usage exceeds the threshold.
 // Works with GPU VRAM if available, otherwise uses system RAM.
@@ -642,7 +649,7 @@ func (o *ApplicationConfig) ToRuntimeSettings() RuntimeSettings {
 		AutoloadBackendGalleries: &autoloadBackendGalleries,
 		ApiKeys:                  &apiKeys,
 		AgentJobRetentionDays:    &agentJobRetentionDays,
-		OpenResponsesStoreTTL:     &openResponsesStoreTTL,
+		OpenResponsesStoreTTL:    &openResponsesStoreTTL,
 	}
 }
 

core/http/app.go

@@ -29,6 +29,8 @@ import (
 //go:embed static/*
 var embedDirStatic embed.FS
 
+var quietPaths = []string{"/api/operations", "/api/resources", "/healthz", "/readyz"}
+
 // @title LocalAI API
 // @version 2.0.0
 // @description The LocalAI Rest API.
@@ -109,10 +111,17 @@ func API(application *application.Application) (*echo.Echo, error) {
 			res := c.Response()
 			err := next(c)
 
-			// Fix for #7989: Reduce log verbosity of Web UI polling and resources API
-			// If the path is /api/operations or /api/resources and the request was successful (200),
-			// we log it at DEBUG level (hidden by default) instead of INFO.
-			if (req.URL.Path == "/api/operations" || req.URL.Path == "/api/resources") && res.Status == 200 {
+			// Fix for #7989: Reduce log verbosity of Web UI polling, resources API, and health checks
+			// These paths are logged at DEBUG level (hidden by default) instead of INFO.
+			isQuietPath := false
+			for _, path := range quietPaths {
+				if req.URL.Path == path {
+					isQuietPath = true
+					break
+				}
+			}
+
+			if isQuietPath && res.Status == 200 {
 				xlog.Debug("HTTP request", "method", req.Method, "path", req.URL.Path, "status", res.Status)
 			} else {
 				xlog.Info("HTTP request", "method", req.Method, "path", req.URL.Path, "status", res.Status)

core/http/endpoints/openai/realtime.go

@@ -1046,6 +1046,27 @@ func triggerResponse(session *Session, conv *Conversation, c *LockedWebsocket, o
 					Content:       content.Text,
 				})
 			}
+		} else if item.FunctionCall != nil {
+			conversationHistory = append(conversationHistory, schema.Message{
+				Role: string(types.MessageRoleAssistant),
+				ToolCalls: []schema.ToolCall{
+					{
+						ID:   item.FunctionCall.CallID,
+						Type: "function",
+						FunctionCall: schema.FunctionCall{
+							Name:      item.FunctionCall.Name,
+							Arguments: item.FunctionCall.Arguments,
+						},
+					},
+				},
+			})
+		} else if item.FunctionCallOutput != nil {
+			conversationHistory = append(conversationHistory, schema.Message{
+				Role:          "tool",
+				Name:          item.FunctionCallOutput.CallID,
+				Content:       item.FunctionCallOutput.Output,
+				StringContent: item.FunctionCallOutput.Output,
+			})
 		}
 	}
 	conv.Lock.Unlock()

core/http/endpoints/openresponses/responses.go

@@ -279,6 +279,18 @@ func convertORInputToMessages(input interface{}, cfg *config.ModelConfig) ([]sch
 					return nil, err
 				}
 				messages = append(messages, msg)
+			case "reasoning":
+				msg, err := convertORReasoningItemToMessage(itemMap)
+				if err != nil {
+					return nil, err
+				}
+				messages = append(messages, msg)
+			case "function_call":
+				msg, err := convertORFunctionCallItemToMessage(itemMap)
+				if err != nil {
+					return nil, err
+				}
+				messages = append(messages, msg)
 			case "function_call_output":
 				// Convert function call output to tool role message
 				callID, _ := itemMap["call_id"].(string)
@@ -323,12 +335,59 @@ func convertORInputToMessages(input interface{}, cfg *config.ModelConfig) ([]sch
 				messages = append(messages, msg)
 			}
 		}
-		return messages, nil
+		return mergeContiguousAssistantMessages(messages), nil
 	default:
 		return nil, fmt.Errorf("unsupported input type: %T", input)
 	}
 }
 
+// convertORReasoningItemToMessage converts an Open Responses reasoning item to an assistant Message fragment (for merging).
+func convertORReasoningItemToMessage(itemMap map[string]interface{}) (schema.Message, error) {
+	var reasoning string
+	if content := itemMap["content"]; content != nil {
+		if s, ok := content.(string); ok {
+			reasoning = s
+		} else if parts, ok := content.([]interface{}); ok {
+			for _, p := range parts {
+				if partMap, ok := p.(map[string]interface{}); ok {
+					if t, _ := partMap["type"].(string); (t == "output_text" || t == "input_text") && partMap["text"] != nil {
+						if tStr, ok := partMap["text"].(string); ok {
+							reasoning += tStr
+						}
+					}
+				}
+			}
+		}
+	}
+	return schema.Message{Role: "assistant", Reasoning: stringPtr(reasoning)}, nil
+}
+
+// convertORFunctionCallItemToMessage converts an Open Responses function_call item to an assistant Message fragment (for merging).
+func convertORFunctionCallItemToMessage(itemMap map[string]interface{}) (schema.Message, error) {
+	callID, _ := itemMap["call_id"].(string)
+	name, _ := itemMap["name"].(string)
+	arguments, _ := itemMap["arguments"].(string)
+	if callID == "" {
+		callID = fmt.Sprintf("call_%s", name)
+	}
+	return schema.Message{
+		Role: "assistant",
+		ToolCalls: []schema.ToolCall{{
+			Index:        0,
+			ID:           callID,
+			Type:         "function",
+			FunctionCall: schema.FunctionCall{Name: name, Arguments: arguments},
+		}},
+	}, nil
+}
+
+func stringPtr(s string) *string {
+	if s == "" {
+		return nil
+	}
+	return &s
+}
+
 // convertORItemToMessage converts a single ORItemField to a Message
 // responseID is the ID of the response where this item was found (for logging/debugging)
 func convertORItemToMessage(item *schema.ORItemField, responseID string) (schema.Message, error) {
@@ -366,19 +425,52 @@ func convertORItemToMessage(item *schema.ORItemField, responseID string) (schema
 			Content:       outputStr,
 			StringContent: outputStr,
 		}, nil
+	case "reasoning":
+		reasoning := extractReasoningContentFromORItem(item)
+		return schema.Message{Role: "assistant", Reasoning: stringPtr(reasoning)}, nil
+	case "function_call":
+		callID := item.CallID
+		if callID == "" {
+			callID = fmt.Sprintf("call_%s", item.Name)
+		}
+		return schema.Message{
+			Role: "assistant",
+			ToolCalls: []schema.ToolCall{{
+				Index:        0,
+				ID:           callID,
+				Type:         "function",
+				FunctionCall: schema.FunctionCall{Name: item.Name, Arguments: item.Arguments},
+			}},
+		}, nil
 	default:
 		return schema.Message{}, fmt.Errorf("unsupported item type for conversion: %s (from response %s)", item.Type, responseID)
 	}
 }
 
-// convertOROutputItemsToMessages converts Open Responses output items to internal Messages
+func extractReasoningContentFromORItem(item *schema.ORItemField) string {
+	if contentParts, ok := item.Content.([]schema.ORContentPart); ok {
+		var s string
+		for _, part := range contentParts {
+			if part.Type == "output_text" || part.Type == "input_text" {
+				s += part.Text
+			}
+		}
+		return s
+	}
+	if s, ok := item.Content.(string); ok {
+		return s
+	}
+	return ""
+}
+
+// convertOROutputItemsToMessages converts Open Responses output items to internal Messages.
+// Contiguous assistant items (message, reasoning, function_call) are merged into a single message.
 func convertOROutputItemsToMessages(outputItems []schema.ORItemField) ([]schema.Message, error) {
 	var messages []schema.Message
 
 	for _, item := range outputItems {
 		switch item.Type {
 		case "message":
-			// Convert message item to assistant message
 			var textContent string
 			if contentParts, ok := item.Content.([]schema.ORContentPart); ok && len(contentParts) > 0 {
 				for _, part := range contentParts {
@@ -392,9 +484,23 @@ func convertOROutputItemsToMessages(outputItems []schema.ORItemField) ([]schema.
 				StringContent: textContent,
 				Content:       textContent,
 			})
+		case "reasoning":
+			reasoning := extractReasoningContentFromORItem(&item)
+			messages = append(messages, schema.Message{Role: "assistant", Reasoning: stringPtr(reasoning)})
 		case "function_call":
-			// Function calls are handled separately - they become tool calls in the next turn
-			// For now, we skip them as they're part of the model's output, not input
+			msg := schema.Message{
+				Role: "assistant",
+				ToolCalls: []schema.ToolCall{{
+					Index:        0,
+					ID:           item.CallID,
+					Type:         "function",
+					FunctionCall: schema.FunctionCall{Name: item.Name, Arguments: item.Arguments},
+				}},
+			}
+			if msg.ToolCalls[0].ID == "" {
+				msg.ToolCalls[0].ID = fmt.Sprintf("call_%s", item.Name)
+			}
+			messages = append(messages, msg)
 		case "function_call_output":
 			// Convert function call output to tool role message
 			var outputStr string
@@ -414,7 +520,74 @@ func convertOROutputItemsToMessages(outputItems []schema.ORItemField) ([]schema.
 		}
 	}
 
-	return messages, nil
+	return mergeContiguousAssistantMessages(messages), nil
+}
+
+// mergeContiguousAssistantMessages merges contiguous assistant messages into one.
+// Many chat templates expect content, reasoning, and tool calls in a single assistant message
+// (see e.g. llama.cpp PR 19773). This avoids creating separate messages per input item.
+func mergeContiguousAssistantMessages(messages []schema.Message) []schema.Message {
+	if len(messages) == 0 {
+		return messages
+	}
+	var out []schema.Message
+	var acc *schema.Message
+	for i := range messages {
+		m := &messages[i]
+		if m.Role != "assistant" {
+			flushAssistantAccumulator(&out, &acc)
+			out = append(out, *m)
+			continue
+		}
+		if acc == nil {
+			acc = &schema.Message{Role: "assistant"}
+		}
+		if m.StringContent != "" {
+			if acc.StringContent != "" {
+				acc.StringContent += "\n" + m.StringContent
+			} else {
+				acc.StringContent = m.StringContent
+			}
+			if acc.Content == nil {
+				acc.Content = m.Content
+			} else if _, ok := m.Content.(string); ok {
+				acc.Content = acc.StringContent
+			}
+		}
+		if m.Reasoning != nil && *m.Reasoning != "" {
+			if acc.Reasoning == nil {
+				acc.Reasoning = m.Reasoning
+			} else {
+				combined := *acc.Reasoning + "\n" + *m.Reasoning
+				acc.Reasoning = &combined
+			}
+		}
+		if len(m.ToolCalls) > 0 {
+			acc.ToolCalls = append(acc.ToolCalls, m.ToolCalls...)
+		}
+	}
+	flushAssistantAccumulator(&out, &acc)
+	return out
+}
+
+func flushAssistantAccumulator(out *[]schema.Message, acc **schema.Message) {
+	if acc == nil || *acc == nil {
+		return
+	}
+	m := *acc
+	if m.StringContent == "" && (m.Reasoning == nil || *m.Reasoning == "") && len(m.ToolCalls) == 0 {
+		*acc = nil
+		return
+	}
+	if m.Content == nil {
+		m.Content = m.StringContent
+	}
+	// Re-index tool calls after merge (each may have been 0)
+	for i := range m.ToolCalls {
+		m.ToolCalls[i].Index = i
+	}
+	*out = append(*out, *m)
+	*acc = nil
 }
 
 // convertORMessageItem converts an Open Responses message item to internal Message

core/http/routes/ui.go

@@ -7,6 +7,7 @@ import (
 	"github.com/mudler/LocalAI/core/http/endpoints/localai"
 	"github.com/mudler/LocalAI/core/http/middleware"
 	"github.com/mudler/LocalAI/core/services"
+	"github.com/mudler/LocalAI/core/trace"
 	"github.com/mudler/LocalAI/internal"
 	"github.com/mudler/LocalAI/pkg/model"
 )
@@ -430,4 +431,13 @@ func RegisterUIRoutes(app *echo.Echo,
 		return c.NoContent(204)
 	})
 
+	app.GET("/api/backend-traces", func(c echo.Context) error {
+		return c.JSON(200, trace.GetBackendTraces())
+	})
+
+	app.POST("/api/backend-traces/clear", func(c echo.Context) error {
+		trace.ClearBackendTraces()
+		return c.NoContent(204)
+	})
+
 }

core/http/views/traces.html

@@ -5,7 +5,7 @@
 <body class="bg-[var(--color-bg-primary)] text-[var(--color-text-primary)]">
 <div class="app-layout">
   {{template "views/partials/navbar" .}}
-  
+
   <main class="main-content">
     <div class="main-content-inner" x-data="tracesApp()" x-init="init()">
 
@@ -40,10 +40,10 @@
         <div class="hero-section">
             <div class="hero-content">
                 <h1 class="hero-title">
-                    API Traces
+                    Traces
                 </h1>
-                <p class="hero-subtitle">View logged API requests and responses</p>
-                <div class="flex flex-wrap justify-center gap-2">
+                <p class="hero-subtitle">View logged API requests, responses, and backend operations</p>
+                <div class="flex flex-wrap justify-center gap-2" x-show="activeTab === 'api'">
                     <button type="button" @click="clearTraces()" class="inline-flex items-center gap-1.5 text-xs text-[var(--color-text-secondary)] hover:text-[var(--color-primary)] bg-transparent hover:bg-[var(--color-primary)]/10 border border-[var(--color-border-subtle)] hover:border-[var(--color-primary)]/30 rounded-md py-1.5 px-2.5 transition-colors">
                         <i class="fas fa-trash text-[10px]"></i>
                         <span>Clear Traces</span>
@@ -53,16 +53,42 @@
                         <span>Export Traces</span>
                     </a>
                 </div>
+                <div class="flex flex-wrap justify-center gap-2" x-show="activeTab === 'backend'">
+                    <button type="button" @click="clearBackendTraces()" class="inline-flex items-center gap-1.5 text-xs text-[var(--color-text-secondary)] hover:text-[var(--color-primary)] bg-transparent hover:bg-[var(--color-primary)]/10 border border-[var(--color-border-subtle)] hover:border-[var(--color-primary)]/30 rounded-md py-1.5 px-2.5 transition-colors">
+                        <i class="fas fa-trash text-[10px]"></i>
+                        <span>Clear Backend Traces</span>
+                    </button>
+                    <a href="/api/backend-traces" download="backend-traces.json" class="inline-flex items-center gap-1.5 text-xs text-[var(--color-text-secondary)] hover:text-[var(--color-primary)] bg-transparent hover:bg-[var(--color-primary)]/10 border border-[var(--color-border-subtle)] hover:border-[var(--color-primary)]/30 rounded-md py-1.5 px-2.5 transition-colors">
+                        <i class="fas fa-download text-[10px]"></i>
+                        <span>Export Backend Traces</span>
+                    </a>
+                </div>
             </div>
         </div>
 
+        <!-- Tab Bar -->
+        <div class="flex border-b border-[var(--color-border-subtle)] mb-6">
+            <button @click="switchTab('api')"
+                    :class="activeTab === 'api' ? 'border-[var(--color-primary)] text-[var(--color-primary)]' : 'border-transparent text-[var(--color-text-secondary)] hover:text-[var(--color-text-primary)]'"
+                    class="px-4 py-2 text-sm font-medium border-b-2 transition-colors">
+                <i class="fas fa-exchange-alt mr-1.5 text-xs"></i>API Traces
+                <span class="ml-1 text-xs opacity-70" x-text="'(' + traces.length + ')'"></span>
+            </button>
+            <button @click="switchTab('backend')"
+                    :class="activeTab === 'backend' ? 'border-[var(--color-primary)] text-[var(--color-primary)]' : 'border-transparent text-[var(--color-text-secondary)] hover:text-[var(--color-text-primary)]'"
+                    class="px-4 py-2 text-sm font-medium border-b-2 transition-colors">
+                <i class="fas fa-cogs mr-1.5 text-xs"></i>Backend Traces
+                <span class="ml-1 text-xs opacity-70" x-text="'(' + backendTraces.length + ')'"></span>
+            </button>
+        </div>
+
         <!-- Tracing Settings -->
         <div class="bg-[var(--color-bg-secondary)] border border-[var(--color-primary)]/20 rounded-lg p-6 mb-8">
             <h2 class="text-xl font-semibold text-[var(--color-text-primary)] mb-4 flex items-center">
                 <i class="fas fa-bug mr-2 text-[var(--color-primary)] text-sm"></i>
                 Tracing Settings
             </h2>
-            <p class="text-xs text-[var(--color-text-secondary)] mb-4">Configure API tracing</p>
+            <p class="text-xs text-[var(--color-text-secondary)] mb-4">Configure API and backend tracing</p>
 
             <div class="space-y-4">
                 <!-- Enable Tracing -->
@@ -103,54 +129,176 @@
             </div>
         </div>
 
-        <!-- Traces Table -->
-        <div class="mt-8">
+        <!-- API Traces Table -->
+        <div class="mt-8" x-show="activeTab === 'api'">
             <div class="overflow-x-auto">
                 <table class="w-full border-collapse">
                     <thead>
                         <tr class="border-b border-[var(--color-bg-secondary)]">
+                            <th class="w-8 p-2"></th>
                             <th class="text-left p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Method</th>
                             <th class="text-left p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Path</th>
                             <th class="text-left p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Status</th>
-                            <th class="text-right p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Actions</th>
                         </tr>
                     </thead>
-                    <tbody>
-                        <template x-for="(trace, index) in traces" :key="index">
-                            <tr class="hover:bg-[var(--color-bg-secondary)]/50 border-b border-[var(--color-bg-secondary)] transition-colors">
+                    <template x-for="(trace, index) in traces" :key="index">
+                        <tbody>
+                            <tr @click="toggleTrace(index)"
+                                class="cursor-pointer hover:bg-[var(--color-bg-secondary)]/50 border-b border-[var(--color-bg-secondary)] transition-colors">
+                                <td class="p-2 w-8 text-center">
+                                    <i class="fas fa-chevron-right text-xs text-[var(--color-text-secondary)] transition-transform duration-200"
+                                       :class="expandedTraces[index] ? 'rotate-90' : ''"></i>
+                                </td>
                                 <td class="p-2" x-text="trace.request.method"></td>
                                 <td class="p-2" x-text="trace.request.path"></td>
                                 <td class="p-2" x-text="trace.response.status"></td>
-                                <td class="p-2 text-right">
-                                    <button @click="showDetails(index)" class="text-[var(--color-primary)]/60 hover:text-[var(--color-primary)] hover:bg-[var(--color-primary)]/10 rounded p-1 transition-colors">
-                                        <i class="fas fa-eye text-xs"></i>
-                                    </button>
+                            </tr>
+                            <tr x-show="expandedTraces[index]">
+                                <td colspan="4" class="p-0">
+                                    <div class="p-4 bg-[var(--color-bg-secondary)]/30 border-b border-[var(--color-bg-secondary)]">
+                                        <div class="grid grid-cols-1 lg:grid-cols-2 gap-4">
+                                            <div>
+                                                <h4 class="text-sm font-semibold text-[var(--color-text-primary)] mb-2">Request Body</h4>
+                                                <pre class="overflow-auto max-h-[70vh] p-3 rounded-lg bg-[var(--color-bg-primary)] border border-[var(--color-border-subtle)] text-xs font-mono text-[var(--color-text-secondary)] whitespace-pre-wrap break-words"
+                                                     x-text="formatTraceBody(trace.request.body)"></pre>
+                                            </div>
+                                            <div>
+                                                <h4 class="text-sm font-semibold text-[var(--color-text-primary)] mb-2">Response Body</h4>
+                                                <pre class="overflow-auto max-h-[70vh] p-3 rounded-lg bg-[var(--color-bg-primary)] border border-[var(--color-border-subtle)] text-xs font-mono text-[var(--color-text-secondary)] whitespace-pre-wrap break-words"
+                                                     x-text="formatTraceBody(trace.response.body)"></pre>
+                                            </div>
+                                        </div>
+                                    </div>
                                 </td>
                             </tr>
-                        </template>
-                    </tbody>
+                        </tbody>
+                    </template>
                 </table>
+                <div x-show="traces.length === 0" class="text-center py-8 text-[var(--color-text-secondary)] text-sm">
+                    No API traces recorded yet.
+                </div>
             </div>
         </div>
 
-        <!-- Details Modal -->
-        <div x-show="selectedTrace !== null" class="fixed inset-0 bg-black/50 flex items-center justify-center z-50" @click="selectedTrace = null">
-            <div class="bg-[var(--color-bg-secondary)] rounded-lg p-6 max-w-4xl w-full max-h-[90vh] overflow-auto" @click.stop>
-                <div class="flex justify-between mb-4">
-                    <h2 class="h3">Trace Details</h2>
-                    <button @click="selectedTrace = null" class="text-[var(--color-text-secondary)] hover:text-[var(--color-text-primary)]">
-                        <i class="fas fa-times"></i>
-                    </button>
-                </div>
-                <div class="grid grid-cols-2 gap-4">
-                    <div>
-                        <h3 class="text-lg font-semibold mb-2">Request Body</h3>
-                        <div id="requestEditor" class="h-96 border border-[var(--color-primary-border)]/20"></div>
-                    </div>
-                    <div>
-                        <h3 class="text-lg font-semibold mb-2">Response Body</h3>
-                        <div id="responseEditor" class="h-96 border border-[var(--color-primary-border)]/20"></div>
-                    </div>
+        <!-- Backend Traces Table -->
+        <div class="mt-8" x-show="activeTab === 'backend'">
+            <div class="overflow-x-auto">
+                <table class="w-full border-collapse">
+                    <thead>
+                        <tr class="border-b border-[var(--color-bg-secondary)]">
+                            <th class="w-8 p-2"></th>
+                            <th class="text-left p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Type</th>
+                            <th class="text-left p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Timestamp</th>
+                            <th class="text-left p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Model</th>
+                            <th class="text-left p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Summary</th>
+                            <th class="text-left p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Duration</th>
+                            <th class="text-left p-2 text-xs font-semibold text-[var(--color-text-secondary)]">Status</th>
+                        </tr>
+                    </thead>
+                    <template x-for="(trace, index) in backendTraces" :key="index">
+                        <tbody>
+                            <tr @click="toggleBackendTrace(index)"
+                                class="cursor-pointer hover:bg-[var(--color-bg-secondary)]/50 border-b border-[var(--color-bg-secondary)] transition-colors">
+                                <td class="p-2 w-8 text-center">
+                                    <i class="fas fa-chevron-right text-xs text-[var(--color-text-secondary)] transition-transform duration-200"
+                                       :class="expandedBackendTraces[index] ? 'rotate-90' : ''"></i>
+                                </td>
+                                <td class="p-2">
+                                    <span class="inline-flex items-center px-2 py-0.5 rounded text-xs font-medium"
+                                          :class="getTypeClass(trace.type)"
+                                          x-text="trace.type"></span>
+                                </td>
+                                <td class="p-2 text-xs text-[var(--color-text-secondary)]" x-text="formatTimestamp(trace.timestamp)"></td>
+                                <td class="p-2 text-sm" x-text="trace.model_name || '-'"></td>
+                                <td class="p-2 text-sm max-w-xs truncate" x-text="trace.summary || '-'"></td>
+                                <td class="p-2 text-xs text-[var(--color-text-secondary)]" x-text="formatDuration(trace.duration)"></td>
+                                <td class="p-2">
+                                    <template x-if="!trace.error">
+                                        <i class="fas fa-check-circle text-green-500 text-xs"></i>
+                                    </template>
+                                    <template x-if="trace.error">
+                                        <i class="fas fa-times-circle text-red-500 text-xs" :title="trace.error"></i>
+                                    </template>
+                                </td>
+                            </tr>
+                            <tr x-show="expandedBackendTraces[index]">
+                                <td colspan="7" class="p-0">
+                                    <div class="p-4 bg-[var(--color-bg-secondary)]/30 border-b border-[var(--color-bg-secondary)]">
+                                        <!-- Header info -->
+                                        <div class="grid grid-cols-2 md:grid-cols-4 gap-3 mb-4">
+                                            <div class="bg-[var(--color-bg-primary)] rounded-lg p-3 border border-[var(--color-border-subtle)]">
+                                                <div class="text-xs text-[var(--color-text-secondary)] mb-1">Type</div>
+                                                <span class="inline-flex items-center px-2 py-0.5 rounded text-xs font-medium"
+                                                      :class="getTypeClass(trace.type)"
+                                                      x-text="trace.type"></span>
+                                            </div>
+                                            <div class="bg-[var(--color-bg-primary)] rounded-lg p-3 border border-[var(--color-border-subtle)]">
+                                                <div class="text-xs text-[var(--color-text-secondary)] mb-1">Model</div>
+                                                <div class="text-sm font-medium" x-text="trace.model_name || '-'"></div>
+                                            </div>
+                                            <div class="bg-[var(--color-bg-primary)] rounded-lg p-3 border border-[var(--color-border-subtle)]">
+                                                <div class="text-xs text-[var(--color-text-secondary)] mb-1">Backend</div>
+                                                <div class="text-sm font-medium" x-text="trace.backend || '-'"></div>
+                                            </div>
+                                            <div class="bg-[var(--color-bg-primary)] rounded-lg p-3 border border-[var(--color-border-subtle)]">
+                                                <div class="text-xs text-[var(--color-text-secondary)] mb-1">Duration</div>
+                                                <div class="text-sm font-medium" x-text="formatDuration(trace.duration)"></div>
+                                            </div>
+                                        </div>
+
+                                        <!-- Error banner -->
+                                        <div x-show="trace.error" class="bg-red-500/10 border border-red-500/30 rounded-lg p-3 mb-4">
+                                            <div class="flex items-center gap-2">
+                                                <i class="fas fa-exclamation-triangle text-red-500 text-sm"></i>
+                                                <span class="text-sm text-red-400" x-text="trace.error"></span>
+                                            </div>
+                                        </div>
+
+                                        <!-- Data fields as nested accordions -->
+                                        <template x-if="trace.data && Object.keys(trace.data).length > 0">
+                                            <div>
+                                                <h4 class="text-sm font-semibold text-[var(--color-text-primary)] mb-2">Data Fields</h4>
+                                                <div class="border border-[var(--color-border-subtle)] rounded-lg overflow-hidden">
+                                                    <template x-for="[key, value] in Object.entries(trace.data)" :key="key">
+                                                        <div class="border-b border-[var(--color-border-subtle)] last:border-b-0">
+                                                            <!-- Field header row -->
+                                                            <div @click="isLargeValue(value) && toggleBackendField(index, key)"
+                                                                 class="flex items-center gap-2 px-3 py-2 hover:bg-[var(--color-bg-primary)]/50 transition-colors"
+                                                                 :class="isLargeValue(value) ? 'cursor-pointer' : ''">
+                                                                <template x-if="isLargeValue(value)">
+                                                                    <i class="fas fa-chevron-right text-[10px] text-[var(--color-text-secondary)] transition-transform duration-200 w-3 flex-shrink-0"
+                                                                       :class="isBackendFieldExpanded(index, key) ? 'rotate-90' : ''"></i>
+                                                                </template>
+                                                                <template x-if="!isLargeValue(value)">
+                                                                    <span class="w-3 flex-shrink-0"></span>
+                                                                </template>
+                                                                <span class="text-sm font-mono text-[var(--color-primary)] flex-shrink-0" x-text="key"></span>
+                                                                <template x-if="!isLargeValue(value)">
+                                                                    <span class="font-mono text-xs text-[var(--color-text-secondary)]" x-text="formatValue(value)"></span>
+                                                                </template>
+                                                                <template x-if="isLargeValue(value) && !isBackendFieldExpanded(index, key)">
+                                                                    <span class="text-xs text-[var(--color-text-secondary)] truncate" x-text="truncateValue(value, 120)"></span>
+                                                                </template>
+                                                            </div>
+                                                            <!-- Expanded field value -->
+                                                            <div x-show="isLargeValue(value) && isBackendFieldExpanded(index, key)"
+                                                                 class="px-3 pb-3">
+                                                                <pre class="overflow-auto max-h-[70vh] p-3 rounded-lg bg-[var(--color-bg-primary)] border border-[var(--color-border-subtle)] text-xs font-mono text-[var(--color-text-secondary)] whitespace-pre-wrap break-words"
+                                                                     x-text="formatLargeValue(value)"></pre>
+                                                            </div>
+                                                        </div>
+                                                    </template>
+                                                </div>
+                                            </div>
+                                        </template>
+                                    </div>
+                                </td>
+                            </tr>
+                        </tbody>
+                    </template>
+                </table>
+                <div x-show="backendTraces.length === 0" class="text-center py-8 text-[var(--color-text-secondary)] text-sm">
+                    No backend traces recorded yet.
                 </div>
             </div>
         </div>
@@ -160,37 +308,43 @@
 
 </div>
 
-<!-- CodeMirror -->
-<link rel="stylesheet" href="static/assets/codemirror.min.css">
-<script src="static/assets/codemirror.min.js"></script>
-<script src="static/assets/javascript.min.js"></script>
-
-<!-- Styles from model-editor -->
-<style>
-.CodeMirror {
-    height: 100% !important;
-    font-family: monospace;
-}
-</style>
-
 <script>
 function tracesApp() {
     return {
+        activeTab: 'api',
         traces: [],
-        selectedTrace: null,
-        requestEditor: null,
-        responseEditor: null,
+        backendTraces: [],
+        expandedTraces: {},
+        expandedBackendTraces: {},
+        expandedBackendFields: {},
         notifications: [],
         settings: {
             enable_tracing: false,
             tracing_max_items: 0
         },
         saving: false,
+        refreshInterval: null,
 
         init() {
             this.loadTracingSettings();
             this.fetchTraces();
-            setInterval(() => this.fetchTraces(), 5000);
+            this.fetchBackendTraces();
+            this.startAutoRefresh();
+        },
+
+        switchTab(tab) {
+            this.activeTab = tab;
+        },
+
+        startAutoRefresh() {
+            if (this.refreshInterval) clearInterval(this.refreshInterval);
+            this.refreshInterval = setInterval(() => {
+                if (this.activeTab === 'api') {
+                    this.fetchTraces();
+                } else {
+                    this.fetchBackendTraces();
+                }
+            }, 5000);
         },
 
         async loadTracingSettings() {
@@ -261,71 +415,143 @@ function tracesApp() {
         },
 
         async fetchTraces() {
-            const response = await fetch('/api/traces');
-            this.traces = await response.json();
-        },
-
-        async clearTraces() {
-            if (confirm('Clear all traces?')) {
-                await fetch('/api/traces/clear', { method: 'POST' });
-                this.traces = [];
+            try {
+                const response = await fetch('/api/traces');
+                this.traces = await response.json();
+            } catch (e) {
+                console.error('Error fetching API traces:', e);
             }
         },
 
-        showDetails(index) {
-            this.selectedTrace = index;
-            this.$nextTick(() => {
-                const trace = this.traces[index];
+        async fetchBackendTraces() {
+            try {
+                const response = await fetch('/api/backend-traces');
+                this.backendTraces = await response.json();
+            } catch (e) {
+                console.error('Error fetching backend traces:', e);
+            }
+        },
 
-                const decodeBase64 = (base64) => {
-                    const binaryString = atob(base64);
-                    const bytes = new Uint8Array(binaryString.length);
-                    for (let i = 0; i < binaryString.length; i++) {
-                        bytes[i] = binaryString.charCodeAt(i);
-                    }
-                    return new TextDecoder().decode(bytes);
-                };
+        async clearTraces() {
+            if (confirm('Clear all API traces?')) {
+                await fetch('/api/traces/clear', { method: 'POST' });
+                this.traces = [];
+                this.expandedTraces = {};
+            }
+        },
 
-                const formatBody = (bodyText) => {
-                    try {
-                        const json = JSON.parse(bodyText);
-                        return JSON.stringify(json, null, 2);
-                    } catch {
-                        return bodyText;
-                    }
-                };
+        async clearBackendTraces() {
+            if (confirm('Clear all backend traces?')) {
+                await fetch('/api/backend-traces/clear', { method: 'POST' });
+                this.backendTraces = [];
+                this.expandedBackendTraces = {};
+                this.expandedBackendFields = {};
+            }
+        },
 
-                const reqBody = formatBody(decodeBase64(trace.request.body));
-                const resBody = formatBody(decodeBase64(trace.response.body));
+        toggleTrace(index) {
+            this.expandedTraces = {
+                ...this.expandedTraces,
+                [index]: !this.expandedTraces[index]
+            };
+        },
 
-                if (!this.requestEditor) {
-                    this.requestEditor = CodeMirror(document.getElementById('requestEditor'), {
-                        value: reqBody,
-                        mode: 'javascript',
-                        json: true,
-                        theme: 'default',
-                        lineNumbers: true,
-                        readOnly: true,
-                        lineWrapping: true
-                    });
-                } else {
-                    this.requestEditor.setValue(reqBody);
+        toggleBackendTrace(index) {
+            this.expandedBackendTraces = {
+                ...this.expandedBackendTraces,
+                [index]: !this.expandedBackendTraces[index]
+            };
+        },
+
+        toggleBackendField(index, key) {
+            const fieldKey = index + '-' + key;
+            this.expandedBackendFields = {
+                ...this.expandedBackendFields,
+                [fieldKey]: !this.expandedBackendFields[fieldKey]
+            };
+        },
+
+        isBackendFieldExpanded(index, key) {
+            return !!this.expandedBackendFields[index + '-' + key];
+        },
+
+        formatTraceBody(body) {
+            try {
+                const binaryString = atob(body);
+                const bytes = new Uint8Array(binaryString.length);
+                for (let i = 0; i < binaryString.length; i++) {
+                    bytes[i] = binaryString.charCodeAt(i);
                 }
-
-                if (!this.responseEditor) {
-                    this.responseEditor = CodeMirror(document.getElementById('responseEditor'), {
-                        value: resBody,
-                        mode: 'javascript',
-                        json: true,
-                        theme: 'default',
-                        lineNumbers: true,
-                        readOnly: true,
-                        lineWrapping: true
-                    });
-                } else {
-                    this.responseEditor.setValue(resBody);
+                const text = new TextDecoder().decode(bytes);
+                try {
+                    return JSON.stringify(JSON.parse(text), null, 2);
+                } catch {
+                    return text;
                 }
-            });
+            } catch {
+                return body || '';
+            }
+        },
+
+        formatLargeValue(value) {
+            if (typeof value === 'string') {
+                try {
+                    return JSON.stringify(JSON.parse(value), null, 2);
+                } catch {
+                    return value;
+                }
+            }
+            if (typeof value === 'object') {
+                return JSON.stringify(value, null, 2);
+            }
+            return String(value);
+        },
+
+        formatTimestamp(ts) {
+            if (!ts) return '-';
+            const d = new Date(ts);
+            return d.toLocaleTimeString() + '.' + String(d.getMilliseconds()).padStart(3, '0');
+        },
+
+        formatDuration(ns) {
+            if (!ns) return '-';
+            const ms = ns / 1000000;
+            if (ms < 1000) return ms.toFixed(1) + 'ms';
+            return (ms / 1000).toFixed(2) + 's';
+        },
+
+        getTypeClass(type) {
+            const classes = {
+                'llm': 'bg-blue-500/20 text-blue-400',
+                'embedding': 'bg-purple-500/20 text-purple-400',
+                'transcription': 'bg-yellow-500/20 text-yellow-400',
+                'image_generation': 'bg-green-500/20 text-green-400',
+                'video_generation': 'bg-pink-500/20 text-pink-400',
+                'tts': 'bg-orange-500/20 text-orange-400',
+                'sound_generation': 'bg-teal-500/20 text-teal-400',
+                'rerank': 'bg-indigo-500/20 text-indigo-400',
+                'tokenize': 'bg-gray-500/20 text-gray-400',
+            };
+            return classes[type] || 'bg-gray-500/20 text-gray-400';
+        },
+
+        isLargeValue(value) {
+            if (typeof value === 'string') return value.length > 120;
+            if (typeof value === 'object') return JSON.stringify(value).length > 120;
+            return false;
+        },
+
+        truncateValue(value, maxLen) {
+            let str = typeof value === 'object' ? JSON.stringify(value) : String(value);
+            if (str.length <= maxLen) return str;
+            return str.substring(0, maxLen) + '...';
+        },
+
+        formatValue(value) {
+            if (value === null || value === undefined) return 'null';
+            if (typeof value === 'boolean') return value ? 'true' : 'false';
+            if (typeof value === 'object') return JSON.stringify(value);
+            return String(value);
         }
     }
 }

core/schema/openai.go

@@ -115,7 +115,7 @@ const (
 	TranscriptionResponseFormatVtt         = TranscriptionResponseFormatType("vtt")
 	TranscriptionResponseFormatLrc         = TranscriptionResponseFormatType("lrc")
 	TranscriptionResponseFormatJson        = TranscriptionResponseFormatType("json")
-	TranscriptionResponseFormatJsonVerbose = TranscriptionResponseFormatType("json_verbose")
+	TranscriptionResponseFormatJsonVerbose = TranscriptionResponseFormatType("verbose_json")
 )
 
 type ChatCompletionResponseFormat struct {

core/trace/backend_trace.go

@@ -0,0 +1,130 @@
+package trace
+
+import (
+	"encoding/json"
+	"sort"
+	"sync"
+	"time"
+
+	"github.com/emirpasic/gods/v2/queues/circularbuffer"
+	"github.com/mudler/LocalAI/core/schema"
+	"github.com/mudler/xlog"
+)
+
+type BackendTraceType string
+
+const (
+	BackendTraceLLM              BackendTraceType = "llm"
+	BackendTraceEmbedding        BackendTraceType = "embedding"
+	BackendTraceTranscription    BackendTraceType = "transcription"
+	BackendTraceImageGeneration  BackendTraceType = "image_generation"
+	BackendTraceVideoGeneration  BackendTraceType = "video_generation"
+	BackendTraceTTS              BackendTraceType = "tts"
+	BackendTraceSoundGeneration  BackendTraceType = "sound_generation"
+	BackendTraceRerank           BackendTraceType = "rerank"
+	BackendTraceTokenize         BackendTraceType = "tokenize"
+)
+
+type BackendTrace struct {
+	Timestamp time.Time        `json:"timestamp"`
+	Duration  time.Duration    `json:"duration"`
+	Type      BackendTraceType `json:"type"`
+	ModelName string           `json:"model_name"`
+	Backend   string           `json:"backend"`
+	Summary   string           `json:"summary"`
+	Error     string           `json:"error,omitempty"`
+	Data      map[string]any   `json:"data"`
+}
+
+var backendTraceBuffer *circularbuffer.Queue[*BackendTrace]
+var backendMu sync.Mutex
+var backendLogChan = make(chan *BackendTrace, 100)
+var backendInitOnce sync.Once
+
+func InitBackendTracingIfEnabled(maxItems int) {
+	backendInitOnce.Do(func() {
+		if maxItems <= 0 {
+			maxItems = 100
+		}
+		backendMu.Lock()
+		backendTraceBuffer = circularbuffer.New[*BackendTrace](maxItems)
+		backendMu.Unlock()
+
+		go func() {
+			for t := range backendLogChan {
+				backendMu.Lock()
+				if backendTraceBuffer != nil {
+					backendTraceBuffer.Enqueue(t)
+				}
+				backendMu.Unlock()
+			}
+		}()
+	})
+}
+
+func RecordBackendTrace(t BackendTrace) {
+	select {
+	case backendLogChan <- &t:
+	default:
+		xlog.Warn("Backend trace channel full, dropping trace")
+	}
+}
+
+func GetBackendTraces() []BackendTrace {
+	backendMu.Lock()
+	if backendTraceBuffer == nil {
+		backendMu.Unlock()
+		return []BackendTrace{}
+	}
+	ptrs := backendTraceBuffer.Values()
+	backendMu.Unlock()
+
+	traces := make([]BackendTrace, len(ptrs))
+	for i, p := range ptrs {
+		traces[i] = *p
+	}
+
+	sort.Slice(traces, func(i, j int) bool {
+		return traces[i].Timestamp.Before(traces[j].Timestamp)
+	})
+
+	return traces
+}
+
+func ClearBackendTraces() {
+	backendMu.Lock()
+	if backendTraceBuffer != nil {
+		backendTraceBuffer.Clear()
+	}
+	backendMu.Unlock()
+}
+
+func GenerateLLMSummary(messages schema.Messages, prompt string) string {
+	if len(messages) > 0 {
+		last := messages[len(messages)-1]
+		text := ""
+		switch content := last.Content.(type) {
+		case string:
+			text = content
+		default:
+			b, err := json.Marshal(content)
+			if err == nil {
+				text = string(b)
+			}
+		}
+		if text != "" {
+			return TruncateString(text, 200)
+		}
+	}
+	if prompt != "" {
+		return TruncateString(prompt, 200)
+	}
+	return ""
+}
+
+func TruncateString(s string, maxLen int) string {
+	if len(s) <= maxLen {
+		return s
+	}
+	return s[:maxLen] + "..."
+}

docs/content/features/GPU-acceleration.md

@@ -60,6 +60,22 @@ diffusers:
   scheduler_type: "k_dpmpp_sde"
 ```
 
+### Multi-GPU Support
+
+For multi-GPU support with diffusers, you need to configure the model with `tensor_parallel_size` set to the number of GPUs you want to use.
+
+```yaml
+name: stable-diffusion-multigpu
+model: stabilityai/stable-diffusion-xl-base-1.0
+backend: diffusers
+parameters:
+  tensor_parallel_size: 2 # Number of GPUs to use
+```
+
+The `tensor_parallel_size` parameter is set in the gRPC proto configuration (in `ModelOptions` message, field 55). When this is set to a value greater than 1, the diffusers backend automatically enables `device_map="auto"` to distribute the model across multiple GPUs.
+
+When using diffusers with multiple GPUs, ensure you have sufficient GPU memory across all devices. The model will be automatically distributed across available GPUs. For optimal performance, use GPUs of the same type and memory capacity.
+
 ## CUDA(NVIDIA) acceleration
 
 ### Requirements

docs/content/features/audio-to-text.md

@@ -57,7 +57,7 @@ Result:
 
 ---
 
-You can also specify the `response_format` parameter to be one of `lrc`, `srt`, `vtt`, `text`, `json` or `json_verbose` (default):
+You can also specify the `response_format` parameter to be one of `lrc`, `srt`, `vtt`, `text`, `json` or `verbose_json` (default):
 ```bash
 ## Send the example audio file to the transcriptions endpoint
 curl http://localhost:8080/v1/audio/transcriptions -H "Content-Type: multipart/form-data" -F file="@$PWD/gb1.ogg" -F model="whisper-1" -F response_format="srt"

docs/content/reference/cli-reference.md

@@ -46,6 +46,7 @@ Complete reference for all LocalAI command-line interface (CLI) parameters and e
 | `--watchdog-idle-timeout` | `15m` | Threshold beyond which an idle backend should be stopped | `$LOCALAI_WATCHDOG_IDLE_TIMEOUT`, `$WATCHDOG_IDLE_TIMEOUT` |
 | `--enable-watchdog-busy` | `false` | Enable watchdog for stopping backends that are busy longer than the watchdog-busy-timeout | `$LOCALAI_WATCHDOG_BUSY`, `$WATCHDOG_BUSY` |
 | `--watchdog-busy-timeout` | `5m` | Threshold beyond which a busy backend should be stopped | `$LOCALAI_WATCHDOG_BUSY_TIMEOUT`, `$WATCHDOG_BUSY_TIMEOUT` |
+| `--watchdog-interval` | `500ms` | Interval between watchdog checks (e.g., `500ms`, `5s`, `1m`) | `$LOCALAI_WATCHDOG_INTERVAL`, `$WATCHDOG_INTERVAL` |
 | `--force-eviction-when-busy` | `false` | Force eviction even when models have active API calls (default: false for safety). **Warning:** Enabling this can interrupt active requests | `$LOCALAI_FORCE_EVICTION_WHEN_BUSY`, `$FORCE_EVICTION_WHEN_BUSY` |
 | `--lru-eviction-max-retries` | `30` | Maximum number of retries when waiting for busy models to become idle before eviction | `$LOCALAI_LRU_EVICTION_MAX_RETRIES`, `$LRU_EVICTION_MAX_RETRIES` |
 | `--lru-eviction-retry-interval` | `1s` | Interval between retries when waiting for busy models to become idle (e.g., `1s`, `2s`) | `$LOCALAI_LRU_EVICTION_RETRY_INTERVAL`, `$LRU_EVICTION_RETRY_INTERVAL` |

docs/data/version.json

@@ -1,3 +1,3 @@
 {
-  "version": "v3.11.0"
+  "version": "v3.12.1"
 }

go.mod

@@ -1,15 +1,13 @@
 module github.com/mudler/LocalAI
 
-go 1.24.4
-
-toolchain go1.24.5
+go 1.25.0
 
 require (
 	dario.cat/mergo v1.0.2
-	fyne.io/fyne/v2 v2.7.2
+	fyne.io/fyne/v2 v2.7.3
 	github.com/Masterminds/sprig/v3 v3.3.0
 	github.com/alecthomas/kong v1.14.0
-	github.com/anthropics/anthropic-sdk-go v1.22.0
+	github.com/anthropics/anthropic-sdk-go v1.26.0
 	github.com/charmbracelet/glamour v0.10.0
 	github.com/containerd/containerd v1.7.30
 	github.com/dhowden/tag v0.0.0-20240417053706-3d75831295e8
@@ -21,7 +19,7 @@ require (
 	github.com/gofrs/flock v0.13.0
 	github.com/google/go-containerregistry v0.20.7
 	github.com/google/uuid v1.6.0
-	github.com/gpustack/gguf-parser-go v0.23.1
+	github.com/gpustack/gguf-parser-go v0.24.0
 	github.com/hpcloud/tail v1.0.0
 	github.com/ipfs/go-log v1.0.5
 	github.com/jaypipes/ghw v0.23.0
@@ -52,7 +50,7 @@ require (
 	github.com/shirou/gopsutil/v3 v3.24.5
 	github.com/streamer45/silero-vad-go v0.2.1
 	github.com/stretchr/testify v1.11.1
-	github.com/swaggo/echo-swagger v1.4.1
+	github.com/swaggo/echo-swagger v1.5.0
 	github.com/swaggo/swag v1.16.6
 	github.com/testcontainers/testcontainers-go v0.40.0
 	github.com/tmc/langchaingo v0.1.14
@@ -66,9 +64,11 @@ require (
 )
 
 require (
-	github.com/ghodss/yaml v1.0.0 // indirect
+	github.com/labstack/echo/v5 v5.0.0 // indirect
 	github.com/labstack/gommon v0.4.2 // indirect
+	github.com/sv-tools/openapi v0.2.1 // indirect
 	github.com/swaggo/files/v2 v2.0.2 // indirect
+	github.com/swaggo/swag/v2 v2.0.0-rc4 // indirect
 	github.com/tidwall/gjson v1.18.0 // indirect
 	github.com/tidwall/match v1.2.0 // indirect
 	github.com/tidwall/pretty v1.2.1 // indirect
@@ -76,6 +76,7 @@ require (
 	github.com/valyala/fasttemplate v1.2.2 // indirect
 	google.golang.org/protobuf v1.36.11 // indirect
 	gopkg.in/yaml.v2 v2.4.0 // indirect
+	sigs.k8s.io/yaml v1.3.0 // indirect
 )
 
 require (
@@ -101,7 +102,7 @@ require (
 	github.com/go-gl/glfw/v3.3/glfw v0.0.0-20240506104042-037f3cc74f2a // indirect
 	github.com/go-task/slim-sprig/v3 v3.0.0 // indirect
 	github.com/go-text/render v0.2.0 // indirect
-	github.com/go-text/typesetting v0.2.1 // indirect
+	github.com/go-text/typesetting v0.3.3 // indirect
 	github.com/godbus/dbus/v5 v5.1.0 // indirect
 	github.com/google/jsonschema-go v0.4.2 // indirect
 	github.com/hack-pad/go-indexeddb v0.3.2 // indirect

go.sum

@@ -8,8 +8,8 @@ dmitri.shuralyov.com/app/changes v0.0.0-20180602232624-0a106ad413e3/go.mod h1:Yl
 dmitri.shuralyov.com/html/belt v0.0.0-20180602232347-f7d459c86be0/go.mod h1:JLBrvjyP0v+ecvNYvCpyZgu5/xkfAUhi6wJj28eUfSU=
 dmitri.shuralyov.com/service/change v0.0.0-20181023043359-a85b471d5412/go.mod h1:a1inKt/atXimZ4Mv927x+r7UpyzRUf4emIoiiSC2TN4=
 dmitri.shuralyov.com/state v0.0.0-20180228185332-28bcc343414c/go.mod h1:0PRwlb0D6DFvNNtx+9ybjezNCa8XF0xaYcETyp6rHWU=
-fyne.io/fyne/v2 v2.7.2 h1:XiNpWkn0PzX43ZCjbb0QYGg1RCxVbugwfVgikWZBCMw=
-fyne.io/fyne/v2 v2.7.2/go.mod h1:PXbqY3mQmJV3J1NRUR2VbVgUUx3vgvhuFJxyjRK/4Ug=
+fyne.io/fyne/v2 v2.7.3 h1:xBT/iYbdnNHONWO38fZMBrVBiJG8rV/Jypmy4tVfRWE=
+fyne.io/fyne/v2 v2.7.3/go.mod h1:gu+dlIcZWSzKZmnrY8Fbnj2Hirabv2ek+AKsfQ2bBlw=
 fyne.io/systray v1.12.0 h1:CA1Kk0e2zwFlxtc02L3QFSiIbxJ/P0n582YrZHT7aTM=
 fyne.io/systray v1.12.0/go.mod h1:RVwqP9nYMo7h5zViCBHri2FgjXF7H2cub7MAq4NSoLs=
 git.apache.org/thrift.git v0.0.0-20180902110319-2566ecd5d999/go.mod h1:fPE2ZNJGynbRyZ4dJvy6G277gSllfV2HJqblrnkyeyg=
@@ -44,8 +44,8 @@ github.com/andybalholm/brotli v1.0.1/go.mod h1:loMXtMfwqflxFJPmdbJO0a3KNoPuLBgiu
 github.com/andybalholm/brotli v1.2.0 h1:ukwgCxwYrmACq68yiUqwIWnGY0cTPox/M94sVwToPjQ=
 github.com/andybalholm/brotli v1.2.0/go.mod h1:rzTDkvFWvIrjDXZHkuS16NPggd91W3kUSvPlQ1pLaKY=
 github.com/anmitsu/go-shlex v0.0.0-20161002113705-648efa622239/go.mod h1:2FmKhYUyUczH0OGQWaF5ceTx0UBShxjsH6f8oGKYe2c=
-github.com/anthropics/anthropic-sdk-go v1.22.0 h1:sgo4Ob5pC5InKCi/5Ukn5t9EjPJ7KTMaKm5beOYt6rM=
-github.com/anthropics/anthropic-sdk-go v1.22.0/go.mod h1:WTz31rIUHUHqai2UslPpw5CwXrQP3geYBioRV4WOLvE=
+github.com/anthropics/anthropic-sdk-go v1.26.0 h1:oUTzFaUpAevfuELAP1sjL6CQJ9HHAfT7CoSYSac11PY=
+github.com/anthropics/anthropic-sdk-go v1.26.0/go.mod h1:qUKmaW+uuPB64iy1l+4kOSvaLqPXnHTTBKH6RVZ7q5Q=
 github.com/aymanbagabas/go-osc52/v2 v2.0.1 h1:HwpRHbFMcZLEVr42D4p7XBqjyuxQH5SMiErDT4WkJ2k=
 github.com/aymanbagabas/go-osc52/v2 v2.0.1/go.mod h1:uYgXzlJ7ZpABp8OJ+exZzJJhRNQ2ASbcXHWsFqH8hp8=
 github.com/aymanbagabas/go-udiff v0.2.0 h1:TK0fH4MteXUDspT88n8CKzvK0X9O2xu9yQjWpi6yML8=
@@ -128,6 +128,8 @@ github.com/distribution/reference v0.6.0 h1:0IXCQ5g4/QMHHkarYzh5l+u8T3t73zM5Qvfr
 github.com/distribution/reference v0.6.0/go.mod h1:BbU0aIcezP1/5jX/8MP0YiH4SdvB5Y4f/wlDRiLyi3E=
 github.com/dlclark/regexp2 v1.11.0 h1:G/nrcoOa7ZXlpoa/91N3X7mM3r8eIlMBBJZvsz/mxKI=
 github.com/dlclark/regexp2 v1.11.0/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
+github.com/dnaeon/go-vcr v1.2.0 h1:zHCHvJYTMh1N7xnV7zf1m1GPBF9Ad0Jk/whtQ1663qI=
+github.com/dnaeon/go-vcr v1.2.0/go.mod h1:R4UdLID7HZT3taECzJs4YgbbH6PIGXB6W/sc5OLb6RQ=
 github.com/docker/cli v29.0.3+incompatible h1:8J+PZIcF2xLd6h5sHPsp5pvvJA+Sr2wGQxHkRl53a1E=
 github.com/docker/cli v29.0.3+incompatible/go.mod h1:JLrzqnKDaYBop7H2jaqPtU4hHvMKP+vjCwu2uszcLI8=
 github.com/docker/distribution v2.8.3+incompatible h1:AtKxIZ36LoNK51+Z6RpzLpddBirtxJnzDrHLEKxTAYk=
@@ -176,7 +178,6 @@ github.com/fyne-io/image v0.1.1 h1:WH0z4H7qfvNUw5l4p3bC1q70sa5+YWVt6HCj7y4VNyA=
 github.com/fyne-io/image v0.1.1/go.mod h1:xrfYBh6yspc+KjkgdZU/ifUC9sPA5Iv7WYUBzQKK7JM=
 github.com/fyne-io/oksvg v0.2.0 h1:mxcGU2dx6nwjJsSA9PCYZDuoAcsZ/OuJlvg/Q9Njfo8=
 github.com/fyne-io/oksvg v0.2.0/go.mod h1:dJ9oEkPiWhnTFNCmRgEze+YNprJF7YRbpjgpWS4kzoI=
-github.com/ghodss/yaml v1.0.0 h1:wQHKEahhL6wmXdzwWG11gIVCkOv05bNOh+Rxn0yngAk=
 github.com/ghodss/yaml v1.0.0/go.mod h1:4dBDuWmgqj2HViK6kFavaiC9ZROes6MMH2rRYeMEF04=
 github.com/gkampitakis/ciinfo v0.3.2 h1:JcuOPk8ZU7nZQjdUhctuhQofk7BGHuIy0c9Ez8BNhXs=
 github.com/gkampitakis/ciinfo v0.3.2/go.mod h1:1NIwaOcFChN4fa/B0hEBdAb6npDlFL8Bwx4dfRLRqAo=
@@ -218,10 +219,10 @@ github.com/go-task/slim-sprig/v3 v3.0.0 h1:sUs3vkvUymDpBKi3qH1YSqBQk9+9D/8M2mN1v
 github.com/go-task/slim-sprig/v3 v3.0.0/go.mod h1:W848ghGpv3Qj3dhTPRyJypKRiqCdHZiAzKg9hl15HA8=
 github.com/go-text/render v0.2.0 h1:LBYoTmp5jYiJ4NPqDc2pz17MLmA3wHw1dZSVGcOdeAc=
 github.com/go-text/render v0.2.0/go.mod h1:CkiqfukRGKJA5vZZISkjSYrcdtgKQWRa2HIzvwNN5SU=
-github.com/go-text/typesetting v0.2.1 h1:x0jMOGyO3d1qFAPI0j4GSsh7M0Q3Ypjzr4+CEVg82V8=
-github.com/go-text/typesetting v0.2.1/go.mod h1:mTOxEwasOFpAMBjEQDhdWRckoLLeI/+qrQeBCTGEt6M=
-github.com/go-text/typesetting-utils v0.0.0-20241103174707-87a29e9e6066 h1:qCuYC+94v2xrb1PoS4NIDe7DGYtLnU2wWiQe9a1B1c0=
-github.com/go-text/typesetting-utils v0.0.0-20241103174707-87a29e9e6066/go.mod h1:DDxDdQEnB70R8owOx3LVpEFvpMK9eeH1o2r0yZhFI9o=
+github.com/go-text/typesetting v0.3.3 h1:ihGNJU9KzdK2QRDy1Bm7FT5RFQoYb+3n3EIhI/4eaQc=
+github.com/go-text/typesetting v0.3.3/go.mod h1:vIRUT25mLQaSh4C8H/lIsKppQz/Gdb8Pu/tNwpi52ts=
+github.com/go-text/typesetting-utils v0.0.0-20250618110550-c820a94c77b8 h1:4KCscI9qYWMGTuz6BpJtbUSRzcBrUSSE0ENMJbNSrFs=
+github.com/go-text/typesetting-utils v0.0.0-20250618110550-c820a94c77b8/go.mod h1:3/62I4La/HBRX9TcTpBj4eipLiwzf+vhI+7whTc9V7o=
 github.com/go-yaml/yaml v2.1.0+incompatible/go.mod h1:w2MrLa16VYP0jy6N7M5kHaCkaLENm+P+Tv+MfurjSw0=
 github.com/goccy/go-yaml v1.18.0 h1:8W7wMFS12Pcas7KU+VVkaiCng+kG8QiFeFwzFb+rwuw=
 github.com/goccy/go-yaml v1.18.0/go.mod h1:XBurs7gK8ATbW4ZPGKgcbrY1Br56PdM69F7LkFRi1kA=
@@ -294,8 +295,8 @@ github.com/gorilla/css v1.0.1 h1:ntNaBIghp6JmvWnxbZKANoLyuXTPZ4cAMlo6RyhlbO8=
 github.com/gorilla/css v1.0.1/go.mod h1:BvnYkspnSzMmwRK+b8/xgNPLiIuNZr6vbZBTPQ2A3b0=
 github.com/gorilla/websocket v1.5.3 h1:saDtZ6Pbx/0u+bgYQ3q96pZgCzfhKXGPqt7kZ72aNNg=
 github.com/gorilla/websocket v1.5.3/go.mod h1:YR8l580nyteQvAITg2hZ9XVh4b55+EU/adAjf1fMHhE=
-github.com/gpustack/gguf-parser-go v0.23.1 h1:0U7DOrsi7ryx2L/dlMy+BSQ5bJV4AuMEIgGBs4RK46A=
-github.com/gpustack/gguf-parser-go v0.23.1/go.mod h1:y4TwTtDqFWTK+xvprOjRUh+dowgU2TKCX37vRKvGiZ0=
+github.com/gpustack/gguf-parser-go v0.24.0 h1:tdJceXYp9e5RhE9RwVYIuUpir72Jz2D68NEtDXkKCKc=
+github.com/gpustack/gguf-parser-go v0.24.0/go.mod h1:y4TwTtDqFWTK+xvprOjRUh+dowgU2TKCX37vRKvGiZ0=
 github.com/gregjones/httpcache v0.0.0-20180305231024-9cad4c3443a7/go.mod h1:FecbI9+v66THATjSRHfNgh1IVFe/9kFxbXtjV0ctIMA=
 github.com/grpc-ecosystem/grpc-gateway v1.5.0/go.mod h1:RSKVYQBd5MCa4OVpNdGskqpgL2+G+NZTnrVHpWWfpdw=
 github.com/grpc-ecosystem/grpc-gateway v1.16.0 h1:gmcG1KaJ57LophUzW0Hy8NmPhnMZb4M0+kPpLofRdBo=
@@ -392,6 +393,8 @@ github.com/kylelemons/godebug v1.1.0 h1:RPNrshWIDI6G2gRW9EHilWtl7Z6Sb1BR0xunSBf0
 github.com/kylelemons/godebug v1.1.0/go.mod h1:9/0rRGxNHcop5bhtWyNeEfOS8JIWk580+fNqagV/RAw=
 github.com/labstack/echo/v4 v4.15.0 h1:hoRTKWcnR5STXZFe9BmYun9AMTNeSbjHi2vtDuADJ24=
 github.com/labstack/echo/v4 v4.15.0/go.mod h1:xmw1clThob0BSVRX1CRQkGQ/vjwcpOMjQZSZa9fKA/c=
+github.com/labstack/echo/v5 v5.0.0 h1:JHKGrI0cbNsNMyKvranuY0C94O4hSM7yc/HtwcV3Na4=
+github.com/labstack/echo/v5 v5.0.0/go.mod h1:SyvlSdObGjRXeQfCCXW/sybkZdOOQZBmpKF0bvALaeo=
 github.com/labstack/gommon v0.4.2 h1:F8qTUNXgG1+6WQmqoUWnz8WiEU60mXVVw0P4ht1WRA0=
 github.com/labstack/gommon v0.4.2/go.mod h1:QlUFxVM+SNXhDL/Z7YhocGIBYOiwB0mXm1+1bAPHPyU=
 github.com/libp2p/go-buffer-pool v0.1.0 h1:oK4mSFcQz7cTQIfqbe4MIj9gLW+mnanjyFtc6cdF0Y8=
@@ -509,8 +512,6 @@ github.com/morikuni/aec v1.0.0/go.mod h1:BbKIizmSmc5MMPqRYbxO4ZU0S0+P200+tUnFx7P
 github.com/mr-tron/base58 v1.1.2/go.mod h1:BinMc/sQntlIE1frQmRFPUoPA1Zkr8VRgBdjWI2mNwc=
 github.com/mr-tron/base58 v1.2.0 h1:T/HDJBh4ZCPbU39/+c3rRvE0uKBQlU27+QI8LJ4t64o=
 github.com/mr-tron/base58 v1.2.0/go.mod h1:BinMc/sQntlIE1frQmRFPUoPA1Zkr8VRgBdjWI2mNwc=
-github.com/mudler/cogito v0.8.2-0.20260214201734-da0d4ceb2b44 h1:joGszpItINnZdoL/0p2077Wz2xnxMGRSRgYN5mS7I4c=
-github.com/mudler/cogito v0.8.2-0.20260214201734-da0d4ceb2b44/go.mod h1:6sfja3lcu2nWRzEc0wwqGNu/eCG3EWgij+8s7xyUeQ4=
 github.com/mudler/cogito v0.9.1-0.20260217143801-bb7f986ed2c7 h1:z3AcM7LbaQb+C955JdSXksHB9B0uWGQpdgl05gJM+9Y=
 github.com/mudler/cogito v0.9.1-0.20260217143801-bb7f986ed2c7/go.mod h1:6sfja3lcu2nWRzEc0wwqGNu/eCG3EWgij+8s7xyUeQ4=
 github.com/mudler/edgevpn v0.31.1 h1:7qegiDWd0kAg6ljhNHxqvp8hbo/6BbzSdbb7/2WZfiY=
@@ -761,12 +762,16 @@ github.com/stretchr/testify v1.8.3/go.mod h1:sz/lmYIOXD/1dqDmKjjqLyZ2RngseejIcXl
 github.com/stretchr/testify v1.8.4/go.mod h1:sz/lmYIOXD/1dqDmKjjqLyZ2RngseejIcXlSw2iwfAo=
 github.com/stretchr/testify v1.11.1 h1:7s2iGBzp5EwR7/aIZr8ao5+dra3wiQyKjjFuvgVKu7U=
 github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD981Nm4B6U=
-github.com/swaggo/echo-swagger v1.4.1 h1:Yf0uPaJWp1uRtDloZALyLnvdBeoEL5Kc7DtnjzO/TUk=
-github.com/swaggo/echo-swagger v1.4.1/go.mod h1:C8bSi+9yH2FLZsnhqMZLIZddpUxZdBYuNHbtaS1Hljc=
+github.com/sv-tools/openapi v0.2.1 h1:ES1tMQMJFGibWndMagvdoo34T1Vllxr1Nlm5wz6b1aA=
+github.com/sv-tools/openapi v0.2.1/go.mod h1:k5VuZamTw1HuiS9p2Wl5YIDWzYnHG6/FgPOSFXLAhGg=
+github.com/swaggo/echo-swagger v1.5.0 h1:nkHxOaBy0SkbJMtMeXZC64KHSa0mJdZFQhVqwEcMres=
+github.com/swaggo/echo-swagger v1.5.0/go.mod h1:TzO363X1ZG/MSbjrG2IX6m65Yd3/zpqh5KM6lPctAhk=
 github.com/swaggo/files/v2 v2.0.2 h1:Bq4tgS/yxLB/3nwOMcul5oLEUKa877Ykgz3CJMVbQKU=
 github.com/swaggo/files/v2 v2.0.2/go.mod h1:TVqetIzZsO9OhHX1Am9sRf9LdrFZqoK49N37KON/jr0=
 github.com/swaggo/swag v1.16.6 h1:qBNcx53ZaX+M5dxVyTrgQ0PJ/ACK+NzhwcbieTt+9yI=
 github.com/swaggo/swag v1.16.6/go.mod h1:ngP2etMK5a0P3QBizic5MEwpRmluJZPHjXcMoj4Xesg=
+github.com/swaggo/swag/v2 v2.0.0-rc4 h1:SZ8cK68gcV6cslwrJMIOqPkJELRwq4gmjvk77MrvHvY=
+github.com/swaggo/swag/v2 v2.0.0-rc4/go.mod h1:Ow7Y8gF16BTCDn8YxZbyKn8FkMLRUHekv1kROJZpbvE=
 github.com/tarm/serial v0.0.0-20180830185346-98f6abe2eb07/go.mod h1:kDXzergiv9cbyO7IOYJZWg1U88JhDg3PB6klq9Hg2pA=
 github.com/testcontainers/testcontainers-go v0.40.0 h1:pSdJYLOVgLE8YdUY2FHQ1Fxu+aMnb6JfVz1mxk7OeMU=
 github.com/testcontainers/testcontainers-go v0.40.0/go.mod h1:FSXV5KQtX2HAMlm7U3APNyLkkap35zNLxukw9oBi/MY=