feat(crispasr): add F5-TTS support and gallery model (#10753)

Link the f5-tts library into the crispasr backend so CrispASR's native
F5-TTS runtime (SWivid F5-TTS, 22-layer DiT flow-matching + built-in Vocos
vocoder) is compiled in. The single self-contained GGUF auto-detects as
f5-tts through the session router, so no explicit backend selector is
needed. Add the f5-tts-crispasr gallery entry (cstr/f5-tts-GGUF) and an
env-gated e2e synthesis spec.

F5-TTS is voice-cloning only and has no baked speaker: it clones from a
reference WAV plus its transcript, supplied via the voice/voice_text
options. The gallery description documents this bring-your-own-reference
requirement.

Verified e2e on the pinned engine (278fb79): the GGUF auto-detects as
f5-tts, the reference voice loads, and synthesis produces a valid 24 kHz
mono WAV.


Assisted-by: Claude:claude-opus-4-8 [Claude Code]

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
Co-authored-by: Ettore Di Giacinto <mudler@localai.io>

backend/go/crispasr/CMakeLists.txt

@@ -17,7 +17,7 @@ target_link_libraries(gocrispasr PRIVATE
     crispasr-lib
     parakeet canary canary_ctc cohere granite_speech granite_nle
     voxtral voxtral4b qwen3_asr qwen3_tts orpheus chatterbox indextts
-    kokoro voxcpm2_tts m2m100 t5_translate wav2vec2-ggml vibevoice
+    kokoro voxcpm2_tts m2m100 t5_translate wav2vec2-ggml vibevoice f5-tts
     silero-lid pyannote-seg funasr paraformer sensevoice
     crisp_audio
     ggml)

backend/go/crispasr/gocrispasr_test.go

@@ -189,5 +189,32 @@ var _ = Describe("CrispASR", func() {
 			Expect(info.Size()).To(BeNumerically(">", 1024),
 				"expected a non-trivial WAV, got %d bytes", info.Size())
 		})
+
+		It("synthesizes with F5-TTS voice cloning (reference WAV + transcript)", func() {
+			// F5-TTS has no baked speaker: it clones from a reference WAV and
+			// its transcript, supplied via the voice/voice_text options. The
+			// spec skips unless all three fixtures are provided.
+			model := os.Getenv("CRISPASR_F5_MODEL_PATH")
+			refWav := os.Getenv("CRISPASR_F5_REF_WAV")
+			refText := os.Getenv("CRISPASR_F5_REF_TEXT")
+			if model == "" || refWav == "" || refText == "" {
+				Skip("set CRISPASR_F5_MODEL_PATH, CRISPASR_F5_REF_WAV and CRISPASR_F5_REF_TEXT to run this spec")
+			}
+			ensureLibLoaded()
+
+			w := &CrispASR{}
+			Expect(w.Load(&pb.ModelOptions{
+				ModelFile: model,
+				Options:   []string{"voice:" + refWav, "voice_text:" + refText},
+			})).To(Succeed())
+
+			dst := filepath.Join(GinkgoT().TempDir(), "f5.wav")
+			Expect(w.TTS(&pb.TTSRequest{Text: "Hello from LocalAI running F5 text to speech.", Dst: dst})).To(Succeed())
+
+			info, err := os.Stat(dst)
+			Expect(err).ToNot(HaveOccurred(), "synthesized WAV should exist at %q", dst)
+			Expect(info.Size()).To(BeNumerically(">", 1024),
+				"expected a non-trivial WAV, got %d bytes", info.Size())
+		})
 	})
 })

gallery/index.yaml

@@ -36811,6 +36811,34 @@
     - filename: snac-24khz.gguf
       uri: huggingface://cstr/snac-24khz-GGUF/snac-24khz.gguf
       sha256: b4b044631df62ececa86ab080516b3e619cd8f93caabd5f6758c7eae14981bd8
+- name: f5-tts-crispasr
+  url: github:mudler/LocalAI/gallery/virtual.yaml@master
+  urls:
+    - https://huggingface.co/SWivid/F5-TTS
+    - https://huggingface.co/cstr/f5-tts-GGUF
+  description: |
+    F5-TTS v1 Base (SWivid, MIT) text-to-speech synthesized through the CrispASR backend. A 22-layer DiT flow-matching model with a built-in Vocos vocoder in a single self-contained GGUF (no separate codec). Auto-detected by CrispASR and runs end-to-end on CPU, producing 24 kHz mono audio.
+
+    F5-TTS is a voice-cloning model with no built-in speaker: you must supply a reference clip and its transcript. Add `voice:<reference.wav>` and `voice_text:<transcript of the reference>` to the model's `options` (paths resolve against the model directory) before synthesizing; without a reference the model cannot generate audio.
+
+    Default GGUF size ~945 MB (f16). Quantization below f16 is not recommended for flow-matching models. Synthesis runs a 32-step ODE solver and is compute-heavy on CPU (expect long generation times without a GPU-enabled CrispASR build).
+  tags:
+    - crispasr
+    - tts
+    - text-to-speech
+    - gguf
+    - voice-cloning
+  overrides:
+    backend: crispasr
+    known_usecases:
+      - tts
+    name: f5-tts-crispasr
+    parameters:
+      model: f5-tts-v1-base-f16.gguf
+  files:
+    - filename: f5-tts-v1-base-f16.gguf
+      uri: huggingface://cstr/f5-tts-GGUF/f5-tts-v1-base-f16.gguf
+      sha256: fdd6ba8baa9e1224072cd0bea795f7962daf00e89b5d8c736fb06b3155c8d1f1
 - name: piper-it_IT-riccardo-x_low-crispasr
   url: github:mudler/LocalAI/gallery/virtual.yaml@master
   urls: