feat: huggingface provider added

.github/workflows/pr-tests.yml

@@ -115,6 +115,7 @@ jobs:
           COHERE_API_KEY: ${{ secrets.COHERE_API_KEY }}
           VERTEX_CREDENTIALS: ${{ secrets.VERTEX_CREDENTIALS }}
           VERTEX_PROJECT_ID: ${{ secrets.VERTEX_PROJECT_ID }}
+          HUGGING_FACE_API_KEY: ${{ secrets.HUGGING_FACE_API_KEY }}
         run: |
           echo "Running tests for PR #${{ github.event.pull_request.number || 'manual run' }}"
           ./.github/workflows/scripts/run-tests.sh

.github/workflows/release-pipeline.yml

@@ -115,6 +115,7 @@ jobs:
           VERTEX_CREDENTIALS: ${{ secrets.VERTEX_CREDENTIALS }}
           VERTEX_PROJECT_ID: ${{ secrets.VERTEX_PROJECT_ID }}
           CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
+          HUGGING_FACE_API_KEY: ${{ secrets.HUGGING_FACE_API_KEY }}
           AWS_S3_BUCKET: ${{ secrets.AWS_S3_BUCKET }}
           AWS_BEDROCK_ROLE_ARN: ${{ secrets.AWS_BEDROCK_ROLE_ARN }}
         run: ./.github/workflows/scripts/release-core.sh "${{ needs.detect-changes.outputs.core-version }}"
@@ -189,6 +190,7 @@ jobs:
           VERTEX_CREDENTIALS: ${{ secrets.VERTEX_CREDENTIALS }}
           VERTEX_PROJECT_ID: ${{ secrets.VERTEX_PROJECT_ID }}
           CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
+          HUGGING_FACE_API_KEY: ${{ secrets.HUGGING_FACE_API_KEY }}
         run: ./.github/workflows/scripts/release-framework.sh "${{ needs.detect-changes.outputs.framework-version }}"
 
   plugins-release:
@@ -265,6 +267,7 @@ jobs:
           VERTEX_CREDENTIALS: ${{ secrets.VERTEX_CREDENTIALS }}
           VERTEX_PROJECT_ID: ${{ secrets.VERTEX_PROJECT_ID }}
           CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
+          HUGGING_FACE_API_KEY: ${{ secrets.HUGGING_FACE_API_KEY }}
         run: ./.github/workflows/scripts/release-all-plugins.sh '${{ needs.detect-changes.outputs.changed-plugins }}'
 
   bifrost-http-release:
@@ -353,6 +356,7 @@ jobs:
           VERTEX_CREDENTIALS: ${{ secrets.VERTEX_CREDENTIALS }}
           VERTEX_PROJECT_ID: ${{ secrets.VERTEX_PROJECT_ID }}
           CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
+          HUGGING_FACE_API_KEY: ${{ secrets.HUGGING_FACE_API_KEY }}
         run: ./.github/workflows/scripts/release-bifrost-http.sh "${{ needs.detect-changes.outputs.transport-version }}"
 
   # Docker build amd64

core/bifrost.go

@@ -23,6 +23,7 @@ import (
 	"github.com/maximhq/bifrost/core/providers/elevenlabs"
 	"github.com/maximhq/bifrost/core/providers/gemini"
 	"github.com/maximhq/bifrost/core/providers/groq"
+	"github.com/maximhq/bifrost/core/providers/huggingface"
 	"github.com/maximhq/bifrost/core/providers/mistral"
 	"github.com/maximhq/bifrost/core/providers/nebius"
 	"github.com/maximhq/bifrost/core/providers/ollama"
@@ -1885,6 +1886,8 @@ func (bifrost *Bifrost) createBaseProvider(providerKey schemas.ModelProvider, co
 		return elevenlabs.NewElevenlabsProvider(config, bifrost.logger), nil
 	case schemas.Nebius:
 		return nebius.NewNebiusProvider(config, bifrost.logger)
+	case schemas.HuggingFace:
+		return huggingface.NewHuggingFaceProvider(config, bifrost.logger), nil
 	default:
 		return nil, fmt.Errorf("unsupported provider: %s", targetProviderKey)
 	}

core/changelog.md

@@ -1,2 +1,3 @@
+- feat(providers): added HuggingFace provider using Inference Provider API, support for chat(with stream also), response(with stream also), TTS and speech synthesis
 - fix(mcp): ensure properties field is always set for tools - [@CryptoFewka](https://github.com/CryptoFewka)
 - fix(perplexity): correct search_domain_filter json tag - [@hnoguchigr](https://github.com/hnoguchigr)

core/internal/testutil/account.go

@@ -111,6 +111,7 @@ func (account *ComprehensiveTestAccount) GetConfiguredProviders() ([]schemas.Mod
 		schemas.Cerebras,
 		schemas.Gemini,
 		schemas.OpenRouter,
+		schemas.HuggingFace,
 		schemas.Nebius,
 		ProviderOpenAICustom,
 	}, nil
@@ -323,6 +324,14 @@ func (account *ComprehensiveTestAccount) GetKeysForProvider(ctx *context.Context
 				UseForBatchAPI: bifrost.Ptr(true),
 			},
 		}, nil
+	case schemas.HuggingFace:
+		return []schemas.Key{
+			{
+				Value:  os.Getenv("HUGGING_FACE_API_KEY"),
+				Models: []string{},
+				Weight: 1.0,
+			},
+		}, nil
 	case schemas.Nebius:
 		return []schemas.Key{
 			{
@@ -577,6 +586,19 @@ func (account *ComprehensiveTestAccount) GetConfigForProvider(providerKey schema
 				BufferSize:  10,
 			},
 		}, nil
+	case schemas.HuggingFace:
+		return &schemas.ProviderConfig{
+			NetworkConfig: schemas.NetworkConfig{
+				DefaultRequestTimeoutInSeconds: 300,
+				MaxRetries:                     10, // HuggingFace can be variable
+				RetryBackoffInitial:            2 * time.Second,
+				RetryBackoffMax:                30 * time.Second,
+			},
+			ConcurrencyAndBufferSize: schemas.ConcurrencyAndBufferSize{
+				Concurrency: Concurrency,
+				BufferSize:  10,
+			},
+		}, nil
 	case schemas.Nebius:
 		return &schemas.ProviderConfig{
 			NetworkConfig: schemas.NetworkConfig{
@@ -995,4 +1017,38 @@ var AllProviderConfigs = []ComprehensiveTestConfig{
 			{Provider: schemas.OpenAI, Model: "gpt-4o-mini"},
 		},
 	},
+	{
+		Provider:             schemas.HuggingFace,
+		ChatModel:            "groq/openai/gpt-oss-120b",
+		VisionModel:          "fireworks-ai/Qwen/Qwen2.5-VL-32B-Instruct",
+		EmbeddingModel:       "sambanova/intfloat/e5-mistral-7b-instruct",
+		TranscriptionModel:   "fal-ai/openai/whisper-large-v3",
+		SpeechSynthesisModel: "fal-ai/hexgrad/Kokoro-82M",
+		Scenarios: TestScenarios{
+			TextCompletion:        false,
+			TextCompletionStream:  false,
+			SimpleChat:            true,
+			CompletionStream:      true,
+			MultiTurnConversation: true,
+			ToolCalls:             true,
+			ToolCallsStreaming:    true,
+			MultipleToolCalls:     false,
+			End2EndToolCalling:    true,
+			AutomaticFunctionCall: true,
+			ImageURL:              true,
+			ImageBase64:           true,
+			MultipleImages:        true,
+			CompleteEnd2End:       true,
+			Embedding:             true,
+			Transcription:         true,
+			TranscriptionStream:   false,
+			SpeechSynthesis:       true,
+			SpeechSynthesisStream: false,
+			Reasoning:             false,
+			ListModels:            true,
+		},
+		Fallbacks: []schemas.Fallback{
+			{Provider: schemas.OpenAI, Model: "gpt-4o-mini"},
+		},
+	},
 }

core/internal/testutil/responses_stream.go

@@ -690,7 +690,7 @@ func RunResponsesStreamTest(t *testing.T, client *bifrost.Bifrost, ctx context.C
 						}
 
 						// Safety check to prevent infinite loops
-						if responseCount > 100 {
+						if responseCount > 300 {
 							goto lifecycleComplete
 						}