mirror of
https://github.com/mudler/LocalAI.git
synced 2026-09-12 22:33:54 -04:00
* fix(advisorylock): set statement_timeout alongside lock_timeout
WithLockCtx already overrides a deployment-wide lock_timeout on its
dedicated connection so a blocking pg_advisory_lock() waits its turn
instead of failing with 55P03. statement_timeout aborts that exact same
statement independently, with SQLSTATE 57014, and was not overridden.
Production roles commonly carry statement_timeout=60s. Any guarded
section longer than that (a cold model load stages for tens of minutes)
therefore killed every concurrent waiter:
advisorylock: acquiring lock 9003261067483446873: ERROR: canceling
statement due to statement timeout (SQLSTATE 57014)
Derive it from the same context budget as lock_timeout, with a matching
RESET so the pooled connection is returned clean.
Assisted-by: Claude Opus 5 [claude-code]
* feat(distributed): add ModelLoadJob, the durable cold-load record
A cold load in distributed mode is a long-running background job, but it
was modelled as a synchronous side effect of an inference request: the
whole of it (backend install, multi-GB staging, checkpoint load) ran
inside the per-model advisory lock. Loading a 35.7 GB GGUF held that lock
for ~20 minutes, so every concurrent request for the same model blocked
on pg_advisory_lock and died at the role's 60s statement_timeout.
Introduce the row that lets the lock shrink to a decision. Exactly one
ModelLoadJob may be active per tracking key; that uniqueness — not the
lifetime of a lock — is what de-duplicates concurrent loaders across
replicas. ClaimLoadJob does its read-then-write under the advisory lock
and nothing else: no network, file or gRPC I/O inside the guarded
section, so a claim costs milliseconds no matter how long the resulting
load takes.
LastProgress is a heartbeat rather than a byte counter. A checkpoint load
legitimately moves zero bytes for many minutes, so a reaper keyed on byte
movement would reclaim a healthy job mid-load; byte progress stays the
concern of load_deadline.go. A job whose heartbeat stops for longer than
the orphan window is reclaimable, so a replica killed mid-load cannot
wedge a model permanently.
Failed jobs keep their row for a short grace so an immediately-following
request reports the real cause instead of silently starting a fresh load
of a model that just failed.
No caller yet — the router moves onto this in the next commit.
Assisted-by: Claude Opus 5 [claude-code]
* refactor(distributed): run cold loads as jobs, outside the advisory lock
Route wrapped the entire cold load — node selection, backend install,
multi-GB staging and the remote LoadModel — in the per-model advisory
lock. The lock's job is to de-duplicate concurrent loaders, a decision
that takes milliseconds; holding it for the tens of minutes the resulting
work takes is what turned a dedup mechanism into a cluster-wide outage
for that model.
Split it into a claim and a run. The claim is the only thing left inside
the lock. The run is a background job owned by the claiming replica and
bounded by the same progress-extended deadline as before; every other
request for that model — local or on another replica — attaches as a
waiter and is served the moment the model is ready, with no duplicate
load and no lock contention.
Waiters share one broadcast rather than an ordered queue: they all want
the identical outcome, so ordering them would add fairness machinery that
changes no result. The local channel wakes same-replica waiters instantly
and a 2s DB poll is the authority, because a waiter on another replica
has no channel to close. On wake a waiter re-runs the warm path rather
than trusting the signal — the model may have been evicted in between.
A waiter whose client disconnects returns immediately and the job keeps
running; it belongs to the job record, not to the request. A failure is
recorded on the row so every waiter reports the real cause, and the row
survives briefly so the next request does not read "no job" as "not
loading" and start a duplicate load of a model that just failed.
The runner heartbeats the row on a fixed interval whether or not bytes
are moving, which is what keeps a legitimately silent checkpoint load
from being reclaimed as an orphan. Phase (installing/staging/loading) and
placement ride to the heartbeat on the context, the same seam
load_deadline.go already uses, so single-host paths are untouched.
Non-distributed mode (no DB) keeps the inline load exactly as it was.
Assisted-by: Claude Opus 5 [claude-code]
* feat(distributed): bound the wait for a loading model and answer with progress
A request whose model is cold-loading now attaches to the running job and
is served the moment the model is ready. That wait has to be bounded: a
held HTTP request cannot survive real infrastructure, and an ingress or LB
idle timeout kills a twenty-minute request regardless of what LocalAI
does.
New LOCALAI_MODEL_LOAD_WAIT (default 60s) bounds the CALLER, never the
load — the job keeps running either way. On expiry the request gets 503
with Retry-After and a structured body naming the model, the node, the
phase, byte progress and an ETA. The `error` envelope keeps OpenAI
clients working; `loading` is additive so they ignore it.
The ETA comes from the job's own observed rate and is omitted rather than
guessed until enough bytes have moved for that rate to mean anything: a
confidently wrong ETA on a twenty-minute wait is worse than none.
Retry-After is that ETA when known, clamped to [5s, 300s], and the wait
budget otherwise.
LOCALAI_MODEL_LOAD_WAIT=0 waits unbounded, for deployments with no proxy
in front. Zero in the config struct still means "unset, use the default",
so the CLI records the operator's zero as ModelLoadWaitUnbounded rather
than losing the distinction.
The distributed branch of ModelLoader.loadModel wrapped the router's
error with %s, which flattened it to a string. Use %w: the typed error is
what the HTTP layer keys the 503 off.
Assisted-by: Claude Opus 5 [claude-code]
* feat(api): add GET /api/models/{id}/load-status
A client that receives 503 while a model stages onto a worker needs
somewhere to poll. This returns the same `loading` object the 503 carries
— phase, node, byte progress and ETA — or 404 when no load is running.
Read-only and observability-shaped, so it is deliberately neither
admin-gated nor feature-gated: it explains a 503 the caller just
received, and hiding that behind a per-modality feature would make the
explanation for a failed image request depend on chat permissions. It
also gets no MCP tool, since there is nothing here an admin would manage
conversationally.
Registered on the surfaces from .agents/api-endpoints-and-auth.md: the
swagger block (existing `models` tag, so /api/instructions needs no new
area), the endpoint discovery maps in RegisterLocalAIRoutes, regenerated
swagger, and the distributed-mode docs page. No FLAG_* usecase is
involved, so capabilities.js is unchanged.
Assisted-by: Claude Opus 5 [claude-code]
* feat(ui): show cold-load progress in Chat and retry when the model is ready
A chat request for a model that is still staging onto a worker now gets a
503 carrying live progress instead of an error. Render it: the composer
shows the phase (installing / staging / loading), the node, the percent
and the ETA, then polls load-status and re-sends the request the moment
the model is ready.
Reuses the staging progress idiom the page already had rather than
inventing a second one — the two sources are folded into one
loadProgress, with the load job winning because it is authoritative
across frontend replicas and knows the phase, where the staging operation
only knows about a byte transfer this replica happens to be performing.
Waiting is bounded (three send attempts, ~30 min of polling each), so a
load that never finishes still surfaces as an error rather than as a
spinner nobody questions. An aborted generation stops the polling too.
Assisted-by: Claude Opus 5 [claude-code]
* fix(distributed): check warm-path cleanup errors
The router moved legacy cleanup calls onto newly linted lines. Report
cleanup failures while preserving the fallback to a cold load.
Assisted-by: Codex:gpt-5 [golangci-lint]
---------
Co-authored-by: Ettore Di Giacinto <mudler@localai.io>
Co-authored-by: localai-org-maint-bot <306269227+localai-org-maint-bot@users.noreply.github.com>
721 lines
28 KiB
Go
721 lines
28 KiB
Go
package http
|
|
|
|
import (
|
|
"embed"
|
|
"errors"
|
|
"fmt"
|
|
"io/fs"
|
|
"math"
|
|
"mime"
|
|
"net/http"
|
|
"os"
|
|
"path/filepath"
|
|
"strconv"
|
|
"strings"
|
|
"time"
|
|
|
|
"github.com/labstack/echo/v4"
|
|
"github.com/labstack/echo/v4/middleware"
|
|
|
|
"github.com/mudler/LocalAI/pkg/model"
|
|
|
|
"github.com/mudler/LocalAI/core/http/auth"
|
|
"github.com/mudler/LocalAI/core/http/endpoints/localai"
|
|
|
|
httpMiddleware "github.com/mudler/LocalAI/core/http/middleware"
|
|
"github.com/mudler/LocalAI/core/http/routes"
|
|
|
|
"github.com/mudler/LocalAI/core/application"
|
|
"github.com/mudler/LocalAI/core/schema"
|
|
"github.com/mudler/LocalAI/core/services/distributed"
|
|
"github.com/mudler/LocalAI/core/services/finetune"
|
|
"github.com/mudler/LocalAI/core/services/galleryop"
|
|
"github.com/mudler/LocalAI/core/services/messaging"
|
|
"github.com/mudler/LocalAI/core/services/nodes"
|
|
"github.com/mudler/LocalAI/core/services/quantization"
|
|
|
|
"github.com/mudler/xlog"
|
|
)
|
|
|
|
// Embed a directory
|
|
//
|
|
//go:embed static/*
|
|
var embedDirStatic embed.FS
|
|
|
|
// Embed React UI build output
|
|
//
|
|
//go:embed react-ui/dist/*
|
|
var reactUI embed.FS
|
|
|
|
var quietPaths = []string{"/api/operations", "/api/resources", "/healthz", "/readyz"}
|
|
|
|
// immutableAssetCacheControl is the Cache-Control served for content-hashed
|
|
// build output. The filename changes whenever the content does, so a one-year
|
|
// TTL plus `immutable` is safe and removes both the re-download and the
|
|
// conditional revalidation round-trip.
|
|
const immutableAssetCacheControl = "public, max-age=31536000, immutable"
|
|
|
|
func defaultBodyLimitSkipper(c echo.Context) bool {
|
|
// Remeshing accepts generated GLBs that routinely exceed the default
|
|
// upload limit. The route has its own tighter, format-specific limit.
|
|
return c.Request().Method == http.MethodPost && c.Path() == "/3d/remesh"
|
|
}
|
|
|
|
// applyModelLoadCooldown maps a ModelLoadCooldownError anywhere in err's chain
|
|
// to HTTP 503 with a Retry-After header (whole seconds, floor 1), so a client
|
|
// polling a model whose load recently failed backs off instead of triggering a
|
|
// fresh backend start. If err is not a cooldown error, code is returned as-is.
|
|
func applyModelLoadCooldown(err error, code int, c echo.Context) int {
|
|
var coolErr *model.ModelLoadCooldownError
|
|
if !errors.As(err, &coolErr) {
|
|
return code
|
|
}
|
|
secs := int(math.Ceil(coolErr.RetryAfter.Seconds()))
|
|
if secs < 1 {
|
|
secs = 1
|
|
}
|
|
c.Response().Header().Set("Retry-After", strconv.Itoa(secs))
|
|
return http.StatusServiceUnavailable
|
|
}
|
|
|
|
// respondModelLoading answers a request whose model is still cold-loading with
|
|
// 503, a Retry-After header and the live `loading` object, reporting true when
|
|
// it handled the error.
|
|
//
|
|
// The distinction from applyModelLoadCooldown matters: a cooldown means "the
|
|
// last load FAILED, back off", this means "the load is progressing, here is how
|
|
// far it got". Both used to look like the same anonymous error, so an operator
|
|
// watching a 35 GB model stage normally onto a new worker saw only failures.
|
|
func respondModelLoading(err error, c echo.Context) bool {
|
|
var loadErr *nodes.ModelLoadingError
|
|
if !errors.As(err, &loadErr) {
|
|
return false
|
|
}
|
|
setModelLoadingRetryAfter(loadErr, c)
|
|
status := loadErr.Status
|
|
if jerr := c.JSON(http.StatusServiceUnavailable, schema.ModelLoadingResponse{
|
|
Error: &schema.APIError{
|
|
Message: loadErr.Error(),
|
|
Code: "model_loading",
|
|
Type: "model_loading",
|
|
},
|
|
Loading: &status,
|
|
}); jerr != nil {
|
|
xlog.Debug("Failed to write model-loading response", "error", jerr)
|
|
}
|
|
return true
|
|
}
|
|
|
|
// applyModelLoading is the body-less half of respondModelLoading, for the
|
|
// opaque-errors handler: status and Retry-After only.
|
|
func applyModelLoading(err error, code int, c echo.Context) int {
|
|
var loadErr *nodes.ModelLoadingError
|
|
if !errors.As(err, &loadErr) {
|
|
return code
|
|
}
|
|
setModelLoadingRetryAfter(loadErr, c)
|
|
return http.StatusServiceUnavailable
|
|
}
|
|
|
|
func setModelLoadingRetryAfter(loadErr *nodes.ModelLoadingError, c echo.Context) {
|
|
secs := int(math.Ceil(loadErr.RetryAfter.Seconds()))
|
|
if secs < 1 {
|
|
secs = 1
|
|
}
|
|
c.Response().Header().Set("Retry-After", strconv.Itoa(secs))
|
|
}
|
|
|
|
// @title LocalAI API
|
|
// @version 2.0.0
|
|
// @description The LocalAI Rest API.
|
|
// @termsOfService
|
|
// @contact.name LocalAI
|
|
// @contact.url https://localai.io
|
|
// @license.name MIT
|
|
// @license.url https://raw.githubusercontent.com/mudler/LocalAI/master/LICENSE
|
|
// @BasePath /
|
|
// @schemes http https
|
|
// @securityDefinitions.apikey BearerAuth
|
|
// @in header
|
|
// @name Authorization
|
|
// @tag.name inference
|
|
// @tag.description Chat completions, text completions, edits, and responses (OpenAI-compatible)
|
|
// @tag.name embeddings
|
|
// @tag.description Vector embeddings (OpenAI-compatible)
|
|
// @tag.name audio
|
|
// @tag.description Text-to-speech, transcription, voice activity detection, sound generation
|
|
// @tag.name images
|
|
// @tag.description Image generation and inpainting
|
|
// @tag.name video
|
|
// @tag.description Video generation from prompts
|
|
// @tag.name detection
|
|
// @tag.description Object detection in images
|
|
// @tag.name tokenize
|
|
// @tag.description Tokenization and token metrics
|
|
// @tag.name models
|
|
// @tag.description Model gallery browsing, installation, deletion, and listing
|
|
// @tag.name backends
|
|
// @tag.description Backend gallery browsing, installation, deletion, and listing
|
|
// @tag.name config
|
|
// @tag.description Model configuration metadata, autocomplete, PATCH updates, VRAM estimation
|
|
// @tag.name monitoring
|
|
// @tag.description Prometheus metrics, backend status, system information
|
|
// @tag.name mcp
|
|
// @tag.description Model Context Protocol — tool-augmented chat with MCP servers
|
|
// @tag.name agent-jobs
|
|
// @tag.description Agent task and job management
|
|
// @tag.name p2p
|
|
// @tag.description Peer-to-peer networking nodes and tokens
|
|
// @tag.name rerank
|
|
// @tag.description Document reranking
|
|
// @tag.name instructions
|
|
// @tag.description API instruction discovery — browse instruction areas and get endpoint guides
|
|
|
|
func API(application *application.Application) (*echo.Echo, error) {
|
|
e := echo.New()
|
|
|
|
// Set body limit
|
|
if application.ApplicationConfig().UploadLimitMB > 0 {
|
|
e.Use(middleware.BodyLimitWithConfig(middleware.BodyLimitConfig{
|
|
Limit: fmt.Sprintf("%dM", application.ApplicationConfig().UploadLimitMB),
|
|
Skipper: defaultBodyLimitSkipper,
|
|
}))
|
|
}
|
|
|
|
// SPA fallback handler, set later when React UI is available
|
|
var spaFallback func(echo.Context) error
|
|
|
|
// Set error handler
|
|
if !application.ApplicationConfig().OpaqueErrors {
|
|
e.HTTPErrorHandler = func(err error, c echo.Context) {
|
|
if respondModelLoading(err, c) {
|
|
return
|
|
}
|
|
code := http.StatusInternalServerError
|
|
var he *echo.HTTPError
|
|
if errors.As(err, &he) {
|
|
code = he.Code
|
|
}
|
|
code = applyModelLoadCooldown(err, code, c)
|
|
|
|
// Handle 404 errors: serve React SPA for HTML requests, JSON otherwise
|
|
if code == http.StatusNotFound {
|
|
if spaFallback != nil {
|
|
accept := c.Request().Header.Get("Accept")
|
|
contentType := c.Request().Header.Get("Content-Type")
|
|
if strings.Contains(accept, "text/html") && !strings.Contains(contentType, "application/json") {
|
|
spaFallback(c)
|
|
return
|
|
}
|
|
}
|
|
notFoundHandler(c)
|
|
return
|
|
}
|
|
|
|
// Send custom error page
|
|
c.JSON(code, schema.ErrorResponse{
|
|
Error: &schema.APIError{Message: err.Error(), Code: code},
|
|
})
|
|
}
|
|
} else {
|
|
e.HTTPErrorHandler = func(err error, c echo.Context) {
|
|
code := http.StatusInternalServerError
|
|
var he *echo.HTTPError
|
|
if errors.As(err, &he) {
|
|
code = he.Code
|
|
}
|
|
code = applyModelLoadCooldown(err, code, c)
|
|
// Opaque errors deliberately withhold the body, so a still-loading
|
|
// model gets the status and Retry-After but no progress detail.
|
|
code = applyModelLoading(err, code, c)
|
|
c.NoContent(code)
|
|
}
|
|
}
|
|
|
|
// Set renderer
|
|
e.Renderer = renderEngine()
|
|
|
|
// Hide banner
|
|
e.HideBanner = true
|
|
e.HidePort = true
|
|
|
|
// Middleware - StripPathPrefix must be registered early as it uses Rewrite which runs before routing
|
|
e.Pre(httpMiddleware.StripPathPrefix())
|
|
|
|
// Stamp the configured external base URL into each request context so
|
|
// middleware.BaseURL can treat it as authoritative for self-referential
|
|
// links. Registered as Pre so it runs before routing and handlers.
|
|
if extBaseURL := application.ApplicationConfig().ExternalBaseURL; extBaseURL != "" {
|
|
e.Pre(func(next echo.HandlerFunc) echo.HandlerFunc {
|
|
return func(c echo.Context) error {
|
|
c.Set("_external_base_url", extBaseURL)
|
|
return next(c)
|
|
}
|
|
})
|
|
}
|
|
|
|
e.Pre(middleware.RemoveTrailingSlash())
|
|
|
|
if application.ApplicationConfig().MachineTag != "" {
|
|
e.Use(func(next echo.HandlerFunc) echo.HandlerFunc {
|
|
return func(c echo.Context) error {
|
|
c.Response().Header().Set("Machine-Tag", application.ApplicationConfig().MachineTag)
|
|
return next(c)
|
|
}
|
|
})
|
|
}
|
|
|
|
// Security headers (CSP, X-Content-Type-Options, X-Frame-Options,
|
|
// Referrer-Policy). Set early so every response — including 404s and
|
|
// errors — picks them up.
|
|
e.Use(httpMiddleware.SecurityHeaders())
|
|
|
|
// Gzip responses. Registered before the tracing and handler middlewares so
|
|
// the response writer it installs sits underneath them: the trace buffer
|
|
// keeps capturing plaintext while the wire carries the compressed bytes.
|
|
if !application.ApplicationConfig().DisableHTTPCompression {
|
|
e.Use(httpMiddleware.Compression(application.ApplicationConfig().HTTPCompressionMinLength))
|
|
}
|
|
|
|
// Custom logger middleware using xlog
|
|
e.Use(func(next echo.HandlerFunc) echo.HandlerFunc {
|
|
return func(c echo.Context) error {
|
|
req := c.Request()
|
|
res := c.Response()
|
|
err := next(c)
|
|
|
|
// Echo's central HTTPErrorHandler runs *after* this middleware
|
|
// returns, so res.Status still reads the default 200 here when a
|
|
// handler returned an error without writing a response. Mirror
|
|
// echo.DefaultHTTPErrorHandler's status derivation so the access
|
|
// log reflects the status the client actually receives — without
|
|
// this, every silent handler error logs as 200.
|
|
status := res.Status
|
|
if err != nil && !res.Committed {
|
|
status = http.StatusInternalServerError
|
|
var he *echo.HTTPError
|
|
if errors.As(err, &he) {
|
|
status = he.Code
|
|
}
|
|
}
|
|
|
|
// 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 && status == 200 {
|
|
xlog.Debug("HTTP request", "method", req.Method, "path", req.URL.Path, "status", status)
|
|
} else {
|
|
xlog.Info("HTTP request", "method", req.Method, "path", req.URL.Path, "status", status)
|
|
}
|
|
return err
|
|
}
|
|
})
|
|
|
|
// Recover middleware
|
|
if !application.ApplicationConfig().Debug {
|
|
e.Use(middleware.Recover())
|
|
}
|
|
|
|
// Metrics middleware. The metric service was created in
|
|
// application.start() so the OTel global provider is set before any
|
|
// counter is registered (the routing-module billing recorder relies
|
|
// on this). We reuse that instance here rather than calling
|
|
// monitoring.NewLocalAIMetricsService a second time, which would
|
|
// create a second provider, second prometheus exporter, and orphan
|
|
// whichever instance lost the SetMeterProvider race.
|
|
if metricsService := application.MetricsService(); metricsService != nil {
|
|
e.Use(localai.LocalAIMetricsAPIMiddleware(metricsService))
|
|
e.Server.RegisterOnShutdown(func() {
|
|
_ = metricsService.Shutdown()
|
|
})
|
|
}
|
|
|
|
// Health Checks should always be exempt from auth, so register these first
|
|
routes.HealthRoutes(e, application.Ready)
|
|
|
|
// Build auth middleware: use the new auth.Middleware when auth is enabled or
|
|
// as a unified replacement for the legacy key-auth middleware.
|
|
authMiddleware := auth.Middleware(application.AuthDB(), application.ApplicationConfig())
|
|
|
|
// Favicon handler
|
|
e.GET("/favicon.svg", func(c echo.Context) error {
|
|
data, err := embedDirStatic.ReadFile("static/favicon.svg")
|
|
if err != nil {
|
|
return c.NoContent(http.StatusNotFound)
|
|
}
|
|
c.Response().Header().Set("Content-Type", "image/svg+xml")
|
|
return c.Blob(http.StatusOK, "image/svg+xml", data)
|
|
})
|
|
|
|
// Static files - use fs.Sub to create a filesystem rooted at "static"
|
|
staticFS, err := fs.Sub(embedDirStatic, "static")
|
|
if err != nil {
|
|
return nil, fmt.Errorf("failed to create static filesystem: %w", err)
|
|
}
|
|
e.StaticFS("/static", staticFS)
|
|
|
|
// Generated content directories
|
|
if application.ApplicationConfig().GeneratedContentDir != "" {
|
|
os.MkdirAll(application.ApplicationConfig().GeneratedContentDir, 0750)
|
|
audioPath := filepath.Join(application.ApplicationConfig().GeneratedContentDir, "audio")
|
|
imagePath := filepath.Join(application.ApplicationConfig().GeneratedContentDir, "images")
|
|
videoPath := filepath.Join(application.ApplicationConfig().GeneratedContentDir, "videos")
|
|
threeDPath := filepath.Join(application.ApplicationConfig().GeneratedContentDir, "3d")
|
|
|
|
os.MkdirAll(audioPath, 0750)
|
|
os.MkdirAll(imagePath, 0750)
|
|
os.MkdirAll(videoPath, 0750)
|
|
_ = os.MkdirAll(threeDPath, 0750)
|
|
|
|
// Go's built-in MIME table has no .glb entry and minimal containers
|
|
// ship no /etc/mime.types, so generated GLBs would otherwise be
|
|
// served as application/octet-stream.
|
|
_ = mime.AddExtensionType(".glb", "model/gltf-binary")
|
|
|
|
e.Static("/generated-audio", audioPath)
|
|
e.Static("/generated-images", imagePath)
|
|
e.Static("/generated-videos", videoPath)
|
|
e.Static("/generated-3d", threeDPath)
|
|
}
|
|
|
|
// Usage recording is initialised in application/startup.go and
|
|
// surfaced via application.StatsRecorder(); routes wire UsageMiddleware
|
|
// against that recorder regardless of auth state.
|
|
|
|
// Auth is applied to _all_ endpoints. Filtering out endpoints to bypass is
|
|
// the role of the exempt-path logic inside the middleware.
|
|
e.Use(authMiddleware)
|
|
|
|
// Feature and model access control (after auth middleware, before routes)
|
|
if application.AuthDB() != nil {
|
|
e.Use(auth.RequireRouteFeature(application.AuthDB()))
|
|
e.Use(auth.RequireModelAccess(application.AuthDB()))
|
|
e.Use(auth.RequireQuota(application.AuthDB()))
|
|
}
|
|
|
|
// CORS middleware. When CORS=true the operator must also specify the
|
|
// allowed origins; an empty allowlist would otherwise let Echo fall back
|
|
// to AllowOrigins=["*"], which is almost never what someone enabling
|
|
// "strict CORS" intended.
|
|
if application.ApplicationConfig().CORS {
|
|
if application.ApplicationConfig().CORSAllowOrigins == "" {
|
|
xlog.Warn("LOCALAI_CORS=true but LOCALAI_CORS_ALLOW_ORIGINS is empty; refusing to register a wildcard CORS policy. Set the allowlist or unset LOCALAI_CORS.")
|
|
} else {
|
|
corsConfig := middleware.CORSConfig{
|
|
AllowOrigins: strings.Split(application.ApplicationConfig().CORSAllowOrigins, ","),
|
|
}
|
|
e.Use(middleware.CORSWithConfig(corsConfig))
|
|
}
|
|
} else {
|
|
e.Use(middleware.CORS())
|
|
}
|
|
|
|
// CSRF middleware (enabled by default, disable with LOCALAI_DISABLE_CSRF=true)
|
|
//
|
|
// Protection relies on Echo's Sec-Fetch-Site header check (supported by all
|
|
// modern browsers). The legacy cookie+token approach is removed because
|
|
// Echo's Sec-Fetch-Site short-circuit never sets the cookie, so the frontend
|
|
// could never read a token to send back.
|
|
if !application.ApplicationConfig().DisableCSRF {
|
|
xlog.Debug("Enabling CSRF middleware (Sec-Fetch-Site mode)")
|
|
e.Use(middleware.CSRFWithConfig(middleware.CSRFConfig{
|
|
Skipper: func(c echo.Context) bool {
|
|
// Skip CSRF for API clients using auth headers (may be cross-origin)
|
|
if c.Request().Header.Get("Authorization") != "" {
|
|
return true
|
|
}
|
|
if c.Request().Header.Get("x-api-key") != "" || c.Request().Header.Get("xi-api-key") != "" {
|
|
return true
|
|
}
|
|
// Skip when Sec-Fetch-Site header is absent (older browsers, reverse
|
|
// proxies that strip the header). The SameSite=Lax cookie attribute
|
|
// provides baseline CSRF protection for these clients.
|
|
if c.Request().Header.Get("Sec-Fetch-Site") == "" {
|
|
return true
|
|
}
|
|
return false
|
|
},
|
|
// Allow same-site requests (subdomains / different ports) in addition
|
|
// to same-origin which Echo already permits by default.
|
|
AllowSecFetchSiteFunc: func(c echo.Context) (bool, error) {
|
|
secFetchSite := c.Request().Header.Get("Sec-Fetch-Site")
|
|
if secFetchSite == "same-site" {
|
|
return true, nil
|
|
}
|
|
// cross-site: block
|
|
return false, nil
|
|
},
|
|
}))
|
|
}
|
|
|
|
// Admin middleware: enforces admin role when auth is enabled, no-op otherwise
|
|
var adminMiddleware echo.MiddlewareFunc
|
|
if application.AuthDB() != nil {
|
|
adminMiddleware = auth.RequireAdmin()
|
|
} else {
|
|
adminMiddleware = auth.NoopMiddleware()
|
|
}
|
|
|
|
// Feature middlewares: per-feature access control
|
|
agentsMw := auth.RequireFeature(application.AuthDB(), auth.FeatureAgents)
|
|
skillsMw := auth.RequireFeature(application.AuthDB(), auth.FeatureSkills)
|
|
collectionsMw := auth.RequireFeature(application.AuthDB(), auth.FeatureCollections)
|
|
mcpJobsMw := auth.RequireFeature(application.AuthDB(), auth.FeatureMCPJobs)
|
|
|
|
requestExtractor := httpMiddleware.NewRequestExtractor(application.ModelConfigLoader(), application.ModelLoader(), application.ApplicationConfig())
|
|
|
|
// Register auth routes (login, callback, API keys, user management)
|
|
routes.RegisterAuthRoutes(e, application)
|
|
|
|
// Register routing-module usage endpoints. Unlike /api/auth/usage
|
|
// these go through the StatsRecorder and work in no-auth single-user
|
|
// mode by attributing requests to the synthetic "local" user.
|
|
routes.RegisterUsageRoutes(e, application)
|
|
routes.RegisterPIIRoutes(e, application)
|
|
routes.RegisterMiddlewareRoutes(e, application)
|
|
|
|
routes.RegisterElevenLabsRoutes(e, requestExtractor, application.ModelConfigLoader(), application.ModelLoader(), application.ApplicationConfig())
|
|
|
|
// Create opcache for tracking UI operations (used by both UI and LocalAI routes)
|
|
var opcache *galleryop.OpCache
|
|
if !application.ApplicationConfig().DisableWebUI {
|
|
opcache = galleryop.NewOpCache(application.GalleryService())
|
|
// In distributed mode, wire the NATS client + gallery store so this
|
|
// replica's OpCache stays in sync with peers — without this the
|
|
// /api/operations endpoint returns whatever this single replica
|
|
// happened to admit, and a load-balanced UI poll alternates between
|
|
// "operation visible" and "operation gone" between replicas.
|
|
if d := application.Distributed(); d != nil {
|
|
opcache.SetMessagingClient(d.Nats)
|
|
if d.DistStores != nil && d.DistStores.Gallery != nil {
|
|
opcache.SetGalleryStore(d.DistStores.Gallery)
|
|
}
|
|
if err := opcache.Start(application.ApplicationConfig().Context); err != nil {
|
|
xlog.Warn("OpCache distributed subscribe failed; running standalone", "error", err)
|
|
}
|
|
}
|
|
}
|
|
|
|
mcpMw := auth.RequireFeature(application.AuthDB(), auth.FeatureMCP)
|
|
routes.RegisterLocalAIRoutes(e, requestExtractor, application.ModelConfigLoader(), application.ModelLoader(), application.ApplicationConfig(), application.GalleryService(), opcache, application.TemplatesEvaluator(), application, adminMiddleware, mcpJobsMw, mcpMw)
|
|
routes.RegisterAgentPoolRoutes(e, application, agentsMw, skillsMw, collectionsMw)
|
|
// Fine-tuning routes
|
|
fineTuningMw := auth.RequireFeature(application.AuthDB(), auth.FeatureFineTuning)
|
|
// In distributed mode pass the shared NATS client + PostgreSQL store so
|
|
// fine-tune jobs stay consistent across replicas (the SyncedMap broadcasts
|
|
// mutations and hydrates from the DB); standalone passes nil for both.
|
|
var ftNats messaging.MessagingClient
|
|
var ftStore *distributed.FineTuneStore
|
|
if d := application.Distributed(); d != nil {
|
|
ftNats = d.Nats
|
|
if d.DistStores != nil && d.DistStores.FineTune != nil {
|
|
ftStore = d.DistStores.FineTune
|
|
}
|
|
}
|
|
ftService := finetune.NewFineTuneService(
|
|
application.ApplicationConfig(),
|
|
application.ModelLoader(),
|
|
application.ModelConfigLoader(),
|
|
ftNats,
|
|
ftStore,
|
|
)
|
|
routes.RegisterFineTuningRoutes(e, ftService, application.ApplicationConfig(), application, fineTuningMw)
|
|
|
|
// Quantization routes
|
|
quantizationMw := auth.RequireFeature(application.AuthDB(), auth.FeatureQuantization)
|
|
// In distributed mode pass the shared NATS client + PostgreSQL store so
|
|
// quantization jobs stay consistent across replicas (the SyncedMap broadcasts
|
|
// mutations and hydrates from the DB); standalone passes nil for both.
|
|
var quantNats messaging.MessagingClient
|
|
var quantStore *distributed.QuantStore
|
|
if d := application.Distributed(); d != nil {
|
|
quantNats = d.Nats
|
|
if d.DistStores != nil && d.DistStores.Quant != nil {
|
|
quantStore = d.DistStores.Quant
|
|
}
|
|
}
|
|
qService := quantization.NewQuantizationService(
|
|
application.ApplicationConfig(),
|
|
application.ModelLoader(),
|
|
application.ModelConfigLoader(),
|
|
quantNats,
|
|
quantStore,
|
|
)
|
|
routes.RegisterQuantizationRoutes(e, qService, application.ApplicationConfig(), application, quantizationMw)
|
|
|
|
// Node management routes (distributed mode)
|
|
distCfg := application.ApplicationConfig().Distributed
|
|
var registry *nodes.NodeRegistry
|
|
var remoteUnloader nodes.NodeCommandSender
|
|
if d := application.Distributed(); d != nil {
|
|
registry = d.Registry
|
|
if d.Router != nil {
|
|
remoteUnloader = d.Router.Unloader()
|
|
}
|
|
}
|
|
natsCfg := distCfg.NatsAuthConfig()
|
|
routes.RegisterNodeSelfServiceRoutes(e, registry, distCfg.RegistrationToken, distCfg.AutoApproveNodes, application.AuthDB(), application.ApplicationConfig().Auth.APIKeyHMACSecret, natsCfg)
|
|
routes.RegisterNodeAdminRoutes(e, registry, remoteUnloader, application.GalleryService(), opcache, application.ApplicationConfig(), adminMiddleware, application.AuthDB(), application.ApplicationConfig().Auth.APIKeyHMACSecret, application.ApplicationConfig().Distributed.RegistrationToken, natsCfg)
|
|
|
|
// Distributed SSE routes (job progress + agent events via NATS)
|
|
if d := application.Distributed(); d != nil {
|
|
if d.Dispatcher != nil {
|
|
e.GET("/api/agent/jobs/:id/progress", d.Dispatcher.SSEHandler(), mcpJobsMw)
|
|
}
|
|
if d.AgentBridge != nil {
|
|
e.GET("/api/agents/:name/sse/distributed", d.AgentBridge.SSEHandler(), agentsMw)
|
|
}
|
|
}
|
|
|
|
routes.RegisterOpenAIRoutes(e, requestExtractor, application)
|
|
routes.RegisterAnthropicRoutes(e, requestExtractor, application)
|
|
routes.RegisterOpenResponsesRoutes(e, requestExtractor, application)
|
|
routes.RegisterOllamaRoutes(e, requestExtractor, application)
|
|
if application.ApplicationConfig().OllamaAPIRootEndpoint {
|
|
routes.RegisterOllamaRootEndpoint(e)
|
|
}
|
|
if !application.ApplicationConfig().DisableWebUI {
|
|
routes.RegisterUIAPIRoutes(e, application.ModelConfigLoader(), application.ModelLoader(), application.ApplicationConfig(), application.GalleryService(), opcache, application, adminMiddleware)
|
|
routes.RegisterUIRoutes(e, application.ModelConfigLoader(), application.ApplicationConfig(), application.GalleryService(), adminMiddleware)
|
|
|
|
// Serve React SPA from / with SPA fallback via 404 handler
|
|
reactFS, fsErr := fs.Sub(reactUI, "react-ui/dist")
|
|
if fsErr != nil {
|
|
xlog.Warn("React UI not available (build with 'make core/http/react-ui/dist')", "error", fsErr)
|
|
} else {
|
|
serveIndex := func(c echo.Context) error {
|
|
indexHTML, err := reactUI.ReadFile("react-ui/dist/index.html")
|
|
if err != nil {
|
|
return c.String(http.StatusNotFound, "React UI not built")
|
|
}
|
|
// index.html names the content-hashed bundles, so it must never
|
|
// be cached: a stale copy pins the browser to the previous
|
|
// deploy's assets.
|
|
c.Response().Header().Set("Cache-Control", "no-cache")
|
|
// Inject <base href> for reverse-proxy support; baseURL comes
|
|
// from attacker-controllable Host / X-Forwarded-Host headers.
|
|
baseURL := httpMiddleware.BaseURL(c)
|
|
if baseURL != "" {
|
|
baseTag := `<base href="` + httpMiddleware.SecureBaseHref(baseURL) + `" />`
|
|
indexHTML = []byte(strings.Replace(string(indexHTML), "<head>", "<head>\n "+baseTag, 1))
|
|
}
|
|
// <base href> only changes how relative URLs resolve; path-absolute
|
|
// URLs (those starting with `/`) still resolve against the origin
|
|
// and would bypass the reverse-proxy prefix. Rewrite the internal
|
|
// path-absolute references emitted by the build so the browser
|
|
// requests them through the proxy under the prefix.
|
|
//
|
|
// HTML-escape the prefix before interpolating it into attributes:
|
|
// BasePathPrefix already gates X-Forwarded-Prefix via
|
|
// SafeForwardedPrefix, but the validator only blocks open-redirect
|
|
// shapes (// prefix, backslashes, control chars), not attribute
|
|
// breakout characters like `"`. Escaping makes this resilient
|
|
// even if the validator ever loosens.
|
|
if prefix := httpMiddleware.BasePathPrefix(c); prefix != "/" {
|
|
safePrefix := httpMiddleware.SecureBaseHref(prefix)
|
|
html := string(indexHTML)
|
|
html = strings.ReplaceAll(html, `="/assets/`, `="`+safePrefix+`assets/`)
|
|
html = strings.ReplaceAll(html, `="/favicon.svg"`, `="`+safePrefix+`favicon.svg"`)
|
|
indexHTML = []byte(html)
|
|
}
|
|
return c.HTMLBlob(http.StatusOK, indexHTML)
|
|
}
|
|
|
|
// Enable SPA fallback in the 404 handler for client-side routing
|
|
spaFallback = serveIndex
|
|
|
|
// Serve React SPA at /app
|
|
e.GET("/app", serveIndex)
|
|
e.GET("/app/*", serveIndex)
|
|
|
|
// prefixRedirect performs a redirect that preserves X-Forwarded-Prefix
|
|
// for reverse-proxy support. The prefix is forgeable on misconfigured
|
|
// proxy chains, so reject anything that isn't a same-origin path.
|
|
prefixRedirect := func(c echo.Context, target string) error {
|
|
if prefix, ok := httpMiddleware.SafeForwardedPrefix(c.Request().Header.Get("X-Forwarded-Prefix")); ok {
|
|
target = strings.TrimSuffix(prefix, "/") + target
|
|
}
|
|
return c.Redirect(http.StatusMovedPermanently, target)
|
|
}
|
|
|
|
// Redirect / to /app
|
|
e.GET("/", func(c echo.Context) error {
|
|
return prefixRedirect(c, "/app")
|
|
})
|
|
|
|
// Backward compatibility: redirect /browse/* to /app/*
|
|
e.GET("/browse", func(c echo.Context) error {
|
|
return prefixRedirect(c, "/app")
|
|
})
|
|
e.GET("/browse/*", func(c echo.Context) error {
|
|
p := c.Param("*")
|
|
return prefixRedirect(c, "/app/"+p)
|
|
})
|
|
|
|
// Serve React static assets (JS, CSS, etc.) and i18n locale JSONs
|
|
// from the embedded React build. cacheControl is stamped on every
|
|
// hit so the browser can reuse the bytes instead of re-fetching
|
|
// ~1.8 MB of bundle on each navigation.
|
|
serveReactSubdir := func(subdir, cacheControl string) echo.HandlerFunc {
|
|
return func(c echo.Context) error {
|
|
p := subdir + "/" + c.Param("*")
|
|
f, err := reactFS.Open(p)
|
|
if err == nil {
|
|
defer f.Close()
|
|
stat, statErr := f.Stat()
|
|
if statErr == nil && !stat.IsDir() {
|
|
contentType := mime.TypeByExtension(filepath.Ext(p))
|
|
if contentType == "" {
|
|
contentType = echo.MIMEOctetStream
|
|
}
|
|
if cacheControl != "" {
|
|
c.Response().Header().Set("Cache-Control", cacheControl)
|
|
}
|
|
return c.Stream(http.StatusOK, contentType, f)
|
|
}
|
|
}
|
|
return echo.NewHTTPError(http.StatusNotFound)
|
|
}
|
|
}
|
|
// Vite content-hashes everything under /assets (Manage-DrwQK63f.js),
|
|
// so a given URL can never change content: cache it for a year and
|
|
// skip revalidation entirely. Locale JSONs keep stable names, so
|
|
// they only get a short TTL.
|
|
e.GET("/assets/*", serveReactSubdir("assets", immutableAssetCacheControl))
|
|
e.GET("/locales/*", serveReactSubdir("locales", "public, max-age=300"))
|
|
}
|
|
}
|
|
routes.RegisterJINARoutes(e, requestExtractor, application.ModelConfigLoader(), application.ModelLoader(), application.ApplicationConfig())
|
|
|
|
// Note: 404 handling is done via HTTPErrorHandler above, no need for catch-all route
|
|
|
|
// HTTP server timeouts.
|
|
//
|
|
// - ReadHeaderTimeout: bounds the slow-headers Slowloris case. 30s is
|
|
// enough for a real client on a poor connection but cuts off a
|
|
// drip-feeding attacker.
|
|
// - IdleTimeout: bounds idle keep-alive connections.
|
|
//
|
|
// We deliberately leave ReadTimeout and WriteTimeout at 0:
|
|
// - Request bodies can be multi-GB model/dataset uploads.
|
|
// - Chat-completion and SSE responses can stream for many minutes.
|
|
// Operators who need stricter limits should front the server with a
|
|
// reverse proxy that terminates slow clients per-request.
|
|
e.Server.ReadHeaderTimeout = 30 * time.Second
|
|
e.Server.IdleTimeout = 120 * time.Second
|
|
|
|
// Log startup message
|
|
e.Server.RegisterOnShutdown(func() {
|
|
xlog.Info("LocalAI API server shutting down")
|
|
})
|
|
|
|
return e, nil
|
|
}
|