# Bifrost: OSS

## OSS

### Overview

- [Bifrost AI Gateway](https://docs.getbifrost.ai/overview.md): The fastest way to build AI applications that never go down. A high-performance AI gateway unifying 20+ providers through a single OpenAI-compatible API.

### Quick Start

- [Bifrost CLI](https://docs.getbifrost.ai/quickstart/cli/getting-started.md): Launch any coding agent through Bifrost with a single command. Automatic configuration, model selection, and MCP integration - no environment variables needed.

#### Gateway

- [Setting Up](https://docs.getbifrost.ai/quickstart/gateway/setting-up.md): Get Bifrost running as an HTTP API gateway in 30 seconds with zero configuration. Perfect for any programming language.
- [Setting up auth](https://docs.getbifrost.ai/quickstart/gateway/setting-up-auth.md): Learn how to enable basic authentication for the Bifrost dashboard to secure your admin interface and API endpoints.
- [Provider Configuration](https://docs.getbifrost.ai/quickstart/gateway/provider-configuration.md): Configure multiple AI providers for custom concurrency, queue sizes, proxy settings, and more.
- [Streaming Responses](https://docs.getbifrost.ai/quickstart/gateway/streaming.md): Receive AI responses in real-time via Server-Sent Events. Perfect for chat applications, audio processing, and real-time transcription where you want immediate results.
- [Tool Calling](https://docs.getbifrost.ai/quickstart/gateway/tool-calling.md): Enable AI models to use external functions and services by defining tool schemas or connecting to Model Context Protocol (MCP) servers. This allows AI to interact with databases, APIs, file systems, and more.
- [Multimodal Support](https://docs.getbifrost.ai/quickstart/gateway/multimodal.md): Process multiple types of content including images, audio, and text with AI models. Bifrost supports vision analysis, image generation, speech synthesis, and audio transcription across various providers.
- [Reranking](https://docs.getbifrost.ai/quickstart/gateway/reranking.md): Reorder documents by relevance to a query using /v1/rerank.
- [Integrations](https://docs.getbifrost.ai/quickstart/gateway/integrations.md): Use Bifrost as a drop-in replacement for existing AI provider SDKs with zero code changes. Just change the base URL and unlock advanced features.

#### Use as Go SDK

- [Setting Up](https://docs.getbifrost.ai/quickstart/go-sdk/setting-up.md): Get Bifrost running in your Go application in 30 seconds with minimal setup and direct code integration.
- [Provider Configuration](https://docs.getbifrost.ai/quickstart/go-sdk/provider-configuration.md): Configure multiple AI providers for custom concurrency, queue sizes, proxy settings, and more.
- [Context Keys](https://docs.getbifrost.ai/quickstart/go-sdk/context-keys.md): Use context keys to configure request behavior, pass metadata, and access response information throughout the request lifecycle.
- [Streaming Responses](https://docs.getbifrost.ai/quickstart/go-sdk/streaming.md): Receive AI responses in real-time as they're generated. Perfect for chat applications, audio processing, and real-time transcription where you want immediate results.
- [Tool Calling](https://docs.getbifrost.ai/quickstart/go-sdk/tool-calling.md): Enable AI models to use external functions and services by defining tool schemas or connecting to Model Context Protocol (MCP) servers. This allows AI to interact with databases, APIs, file systems, and more.
- [Multimodal Support](https://docs.getbifrost.ai/quickstart/go-sdk/multimodal.md): Process multiple types of content including images, audio, and text with AI models. Bifrost supports vision analysis, image generation, speech synthesis, and audio transcription across various providers.
- [Reranking](https://docs.getbifrost.ai/quickstart/go-sdk/reranking.md): Rerank documents with Bifrost Go SDK using client.RerankRequest.
- [Logging](https://docs.getbifrost.ai/quickstart/go-sdk/logger.md): Configure logging for debugging, monitoring, and troubleshooting your Bifrost integration.

### Release Cadence

- [Release Cadence](https://docs.getbifrost.ai/release-cadence.md): How Bifrost OSS releases are versioned and shipped

### Migration Guides

- [Migrating to v2.0.0](https://docs.getbifrost.ai/migration-guides/v2.0.0.md): Breaking changes and migration instructions for the v2.0.0 release
- [Migrating to v1.5.0](https://docs.getbifrost.ai/migration-guides/v1.5.0.md): Breaking changes and migration instructions for the v1.5.0 release

### Migrate from other Gateway

- [Migrating from LiteLLM](https://docs.getbifrost.ai/migration-guides/litellm.md): Automatically migrate models, keys, organizations, teams, users, and virtual keys from a LiteLLM proxy to Bifrost.

### SDK Integrations

- [What is an integration?](https://docs.getbifrost.ai/integrations/what-is-an-integration.md): Protocol adapters that translate between Bifrost's unified API and provider-specific API formats like OpenAI, Anthropic, and Google GenAI.

### Providers & Guides

- [Routing Rules](https://docs.getbifrost.ai/providers/routing-rules.md): Configure dynamic, expression-based routing decisions using CEL expressions to control how requests are routed across providers.
- [Provider Routing](https://docs.getbifrost.ai/providers/provider-routing.md): Understand how Bifrost routes requests across AI providers using governance rules and adaptive load balancing.
- [Aliasing Models](https://docs.getbifrost.ai/providers/aliasing-models.md): Map arbitrary model names to any target identifier using static key-level aliases or dynamic routing rules.
- [Reasoning](https://docs.getbifrost.ai/providers/reasoning.md): Cross-provider reference for reasoning and thinking capabilities in AI models
- [Performance Tuning](https://docs.getbifrost.ai/providers/performance.md): Optimize Bifrost for high throughput with concurrency, buffer sizing, and memory pool configuration
- [Custom Providers](https://docs.getbifrost.ai/providers/custom-providers.md): Create custom provider configurations with specific request type restrictions, custom naming, and controlled access patterns.
- [Custom Pricing](https://docs.getbifrost.ai/providers/custom-pricing.md): Set custom rates for any model across global or virtual key scopes, optionally narrowed to a specific provider or key.
- [Request Options](https://docs.getbifrost.ai/providers/request-options.md): Complete reference of all request options supported by Bifrost, including HTTP headers for the gateway and context keys for the Go SDK.
- [Test Harness Coverage](https://docs.getbifrost.ai/providers/test-harness-coverage.md): Per-provider matrix of features exercised by Bifrost's end-to-end Postman/newman test harness.

#### Supported Providers

- [Overview](https://docs.getbifrost.ai/providers/supported-providers/overview.md): Bifrost supports multiple AI providers with consistent OpenAI-compatible response formats, enabling seamless provider switching without code changes.
- [Anthropic](https://docs.getbifrost.ai/providers/supported-providers/anthropic.md): Anthropic API conversion guide - structural differences, message handling, thinking/reasoning, and tool conversion
- [Azure](https://docs.getbifrost.ai/providers/supported-providers/azure.md): Azure OpenAI Service API conversion guide - deployment management, authentication, multi-model support
- [Azure Model Router](https://docs.getbifrost.ai/providers/supported-providers/azure-model-router.md): Use Azure model-router deployments with Bifrost - native Chat Completions support and a custom-provider workaround for Responses
- [AWS Bedrock](https://docs.getbifrost.ai/providers/supported-providers/bedrock.md): AWS Bedrock API conversion guide - model families, parameter mapping, message handling, reasoning/thinking, tool conversion, and AWS authentication
- [AWS Bedrock Mantle](https://docs.getbifrost.ai/providers/supported-providers/bedrock-mantle.md): AWS Bedrock Mantle provider - a single endpoint serving Claude (native Anthropic Messages) and OpenAI-family / Gemma models (OpenAI-compatible), with AWS SigV4 or API-key auth
- [Cerebras](https://docs.getbifrost.ai/providers/supported-providers/cerebras.md): Cerebras API conversion guide - OpenAI-compatible format, full feature support, streaming, tool calling, and parameter handling
- [Cohere](https://docs.getbifrost.ai/providers/supported-providers/cohere.md): Cohere API conversion guide - parameter mapping, message handling, reasoning/thinking, and tool conversion
- [Databricks](https://docs.getbifrost.ai/providers/supported-providers/databricks.md): Databricks Model Serving (Foundation Model APIs) and Unity AI Gateway model services - chat, streaming, embeddings, Responses API, PAT and OAuth M2M authentication
- [DeepSeek](https://docs.getbifrost.ai/providers/supported-providers/deepseek.md): DeepSeek API conversion guide - OpenAI-compatible chat and responses by default, optional per-key/per-alias Anthropic-compatible endpoints, OpenAI-compatible beta text completions, streaming, tool calling, and reasoning
- [ElevenLabs](https://docs.getbifrost.ai/providers/supported-providers/elevenlabs.md): ElevenLabs API conversion guide - text-to-speech, speech-to-text, voice settings, and model management
- [Fireworks](https://docs.getbifrost.ai/providers/supported-providers/fireworks.md): Fireworks API conversion guide covering native chat, responses, completions, embeddings, streaming, and Fireworks-specific parameter handling
- [Google Gemini](https://docs.getbifrost.ai/providers/supported-providers/gemini.md): Google Gemini API conversion guide - request/response transformation, message conversion, tool handling, and streaming behavior
- [GitHub Copilot](https://docs.getbifrost.ai/providers/supported-providers/github-copilot.md): Route requests to GitHub Copilot through Bifrost, billed to your organization
- [Groq](https://docs.getbifrost.ai/providers/supported-providers/groq.md): Groq API conversion guide - OpenAI-compatible format, parameter handling, text completion fallback, streaming, and tool support
- [Hugging Face](https://docs.getbifrost.ai/providers/supported-providers/huggingface.md): Detailed guide on Hugging Face provider implementation specifics, including model aliases and unique request handling.
- [Mistral](https://docs.getbifrost.ai/providers/supported-providers/mistral.md): Mistral API conversion guide - parameter mapping, message handling, tool support, transcription, OCR, and streaming behavior
- [Nebius](https://docs.getbifrost.ai/providers/supported-providers/nebius.md): Nebius API conversion guide - OpenAI-compatible format, parameter handling, streaming, embeddings, and special features
- [Ollama](https://docs.getbifrost.ai/providers/supported-providers/ollama.md): Ollama API conversion guide - local inference, OpenAI-compatible format, streaming, tool calling, and embeddings
- [OpenCode](https://docs.getbifrost.ai/providers/supported-providers/opencode.md): OpenCode Zen and Go API gateway providers — OpenAI-compatible format, multi-model access, pay-as-you-go (Zen) and subscription (Go) gateways
- [OpenAI](https://docs.getbifrost.ai/providers/supported-providers/openai.md): OpenAI API conversion guide - what to know when using OpenAI through Bifrost
- [OpenRouter](https://docs.getbifrost.ai/providers/supported-providers/openrouter.md): OpenRouter API conversion guide - routing to multiple providers, reasoning support, parameter handling, and streaming
- [Parasail](https://docs.getbifrost.ai/providers/supported-providers/parasail.md): Parasail API conversion guide - OpenAI-compatible format, streaming support, tool calling, and parameter handling
- [Perplexity](https://docs.getbifrost.ai/providers/supported-providers/perplexity.md): Perplexity API conversion guide - OpenAI-compatible with web search integration, parameter mapping, and reasoning support
- [Replicate](https://docs.getbifrost.ai/providers/supported-providers/replicate.md): Replicate API conversion guide - prediction-based architecture, model-specific parameters, and async/sync modes
- [Runware](https://docs.getbifrost.ai/providers/supported-providers/runware.md): Runware API conversion guide - text-to-image, image editing, upscaling, background removal, masking, text/image-to-video, 3D generation, and raw task passthrough
- [Runway ML](https://docs.getbifrost.ai/providers/supported-providers/runway.md): Runway ML API conversion guide - text-to-video, image-to-video, and video-to-video generation
- [Sarvam AI](https://docs.getbifrost.ai/providers/supported-providers/sarvam.md): Sarvam AI conversion guide - OpenAI-compatible chat plus Indic text-to-speech (Bulbul) and speech-to-text (Saaras/Saarika) with custom mapping
- [SGLang](https://docs.getbifrost.ai/providers/supported-providers/sgl.md): SGL/SGLang API conversion guide - OpenAI-compatible format, parameter handling, streaming, tool support
- [Vertex AI](https://docs.getbifrost.ai/providers/supported-providers/vertex.md): Google Vertex AI API conversion guide - multi-model support, OAuth2 authentication, project/region configuration
- [vLLM](https://docs.getbifrost.ai/providers/supported-providers/vllm.md): vLLM API guide - OpenAI- and Anthropic-compatible self-hosted inference, native Responses, chat, embeddings, rerank, transcription, and streaming
- [Wafer](https://docs.getbifrost.ai/providers/supported-providers/wafer.md): Wafer AI conversion guide - OpenAI-compatible chat, streaming, tool calling, reasoning, text completions, and file uploads referenced by ID
- [xAI](https://docs.getbifrost.ai/providers/supported-providers/xai.md): xAI API conversion guide - OpenAI-compatible format, Grok models, vision support, reasoning, and parameter handling

### MCP Gateway

- [Overview](https://docs.getbifrost.ai/mcp/overview.md): Enable AI models to discover and execute external tools dynamically. Transform static chat models into action-capable agents.
- [Connecting to MCP Servers](https://docs.getbifrost.ai/mcp/connecting-to-servers.md): Connect Bifrost to external MCP servers via STDIO, HTTP, or SSE protocols.
- [Connections, States & Lifecycles](https://docs.getbifrost.ai/mcp/connections.md): How Bifrost holds a connection to each MCP server type, every connection state, and what a client's lifecycle looks like end to end.
- [Virtual MCPs](https://docs.getbifrost.ai/mcp/virtual-mcps.md): Bundle tools from one or more MCP servers into a single addressable MCP endpoint, assignable to virtual keys.
- [MCP Sessions](https://docs.getbifrost.ai/mcp/sessions.md): Inspect, re-authenticate, edit, and revoke per-user MCP credentials — both OAuth tokens and submitted header values.
- [Tool Execution](https://docs.getbifrost.ai/mcp/tool-execution.md): Execute MCP tools with full control over approval and conversation flow.
- [Agent Mode (Auto-Execution)](https://docs.getbifrost.ai/mcp/agent-mode.md): Enable autonomous tool execution with configurable auto-approval for building AI agents.
- [Code Mode](https://docs.getbifrost.ai/mcp/code-mode.md): AI writes Python to orchestrate tools. Reduces input token usage by up to 92.8% when using multiple MCP servers.
- [Gateway Authentication](https://docs.getbifrost.ai/mcp/gateway-auth.md): How MCP clients authenticate to Bifrost's /mcp endpoint — virtual key headers or browser-based OAuth 2.1.
- [Tool Hosting](https://docs.getbifrost.ai/mcp/tool-hosting.md): Register custom tools directly in your Go application without external MCP servers.
- [Tool Filtering](https://docs.getbifrost.ai/mcp/filtering.md): Control which MCP tools are available at the client, request, and virtual key levels.

#### Authentication

- [MCP Authentication](https://docs.getbifrost.ai/mcp/auth/overview.md): Pick the right auth type for your MCP servers. Bifrost supports None, Headers, OAuth 2.0, Per-User OAuth, Per-User Headers, and Token Exchange.
- [No Authentication](https://docs.getbifrost.ai/mcp/auth/none.md): Connect to MCP servers that don't require any upstream credential.
- [Header-Based Authentication](https://docs.getbifrost.ai/mcp/auth/headers.md): Configure static HTTP headers (API keys, bearer tokens, custom auth) shared across all requests to an MCP server.
- [Per-User Headers](https://docs.getbifrost.ai/mcp/auth/per-user-headers.md): Each end-user submits their own HTTP headers (API key, signed token, tenant ID) for the MCP server. Lazy-auth model, no upstream OAuth required.
- [OAuth 2.0 Authentication](https://docs.getbifrost.ai/mcp/auth/oauth.md): Admin-side OAuth 2.0 for MCP servers. Single shared token, automatic refresh, PKCE, dynamic client registration.
- [Per-User OAuth](https://docs.getbifrost.ai/mcp/auth/per-user-oauth.md): Each end-user authenticates with the upstream MCP service under their own credentials. Same lazy-auth model on the MCP Gateway and the LLM Gateway.
- [Token Exchange (On-Behalf-Of)](https://docs.getbifrost.ai/mcp/auth/token-exchange.md): Each caller's identity-provider token is exchanged for a short-lived token scoped to the upstream MCP server — delegated, per-call, no shared credential.

### Custom plugins

- [Getting Started](https://docs.getbifrost.ai/plugins/getting-started.md): Learn how to extend Bifrost's functionality by creating custom plugins that intercept and modify requests and responses.
- [Building Dynamically Linked Bifrost Binary](https://docs.getbifrost.ai/plugins/building-dynamic-binary.md): Learn how to build a dynamically linked Bifrost binary required for custom plugin support
- [Plugin Sequencing](https://docs.getbifrost.ai/plugins/sequencing.md): Control the execution order of custom plugins relative to Bifrost's built-in plugins using placement groups and ordering.
- [Plugin Migration Guide](https://docs.getbifrost.ai/plugins/migration-guide.md): How to migrate your Bifrost plugins across HTTP transport hook changes

#### Writing Plugins

- [Writing Go Plugins](https://docs.getbifrost.ai/plugins/writing-go-plugin.md): Step-by-step guide to creating native Go plugins for Bifrost using shared object (.so) files
- [Writing WASM Plugins (Deprecated)](https://docs.getbifrost.ai/plugins/writing-wasm-plugin.md): Deprecated WebAssembly plugin guide. Use native Go plugins today; webhook-based plugins are planned as the next cross-language extension path.

### Open Source Features

- [Drop-in Replacement](https://docs.getbifrost.ai/features/drop-in-replacement.md): Replace your existing AI SDK connections with Bifrost by changing just the base URL. Keep your code, gain advanced features like fallbacks, load balancing, and governance.
- [Retries & Fallbacks](https://docs.getbifrost.ai/features/retries-and-fallbacks.md): Automatic retry with exponential backoff and provider failover. Retries handle transient errors within a provider; fallbacks switch to a different provider when all retries are exhausted.
- [Compatibility Plugin](https://docs.getbifrost.ai/features/compat-plugin.md): Request and response transformations for LiteLLM proxy/SDK compatibility.
- [Load Balance](https://docs.getbifrost.ai/features/keys-management.md): Intelligent API key management with weighted load balancing, model-specific filtering, and automatic failover. Distribute traffic across multiple keys for optimal performance and reliability.
- [Async Inference](https://docs.getbifrost.ai/features/async-inference.md): Submit inference requests asynchronously and poll for results later.
- [Webhooks](https://docs.getbifrost.ai/features/webhooks.md): Get a signed callback when an async inference job finishes, instead of polling for it.
- [Telemetry](https://docs.getbifrost.ai/features/telemetry.md): Comprehensive Prometheus-based monitoring for Bifrost Gateway with custom metrics and labels.
- [OpenTelemetry](https://docs.getbifrost.ai/features/otel.md): Native OpenTelemetry tracing for every LLM call routed through Bifrost — export to any OTLP collector.
- [Content Logging](https://docs.getbifrost.ai/features/observability/content-logging.md): Control which parts of a request and response are persisted, where they land, and who can read them back
- [Latency & Overhead Breakdown](https://docs.getbifrost.ai/features/observability/latency-breakdown.md): Understand where every millisecond of a request goes. A reference for the overhead breakdown shown in the log detail view, with every span bucket explained.
- [Semantic Caching](https://docs.getbifrost.ai/features/semantic-caching.md): Cache AI responses with exact-match hashing and semantic similarity search. Cut costs and latency by replaying answers for identical or semantically similar requests.
- [Auto Prompt Caching](https://docs.getbifrost.ai/features/prompt-caching.md): Automatically inject prompt-cache breakpoints for clients that send none, so agentic tools stop paying the cache-write rate on every turn.
- [Skills Repository](https://docs.getbifrost.ai/features/skills-repository.md): Create, version, and publish Agent Skills from Bifrost, then install them in Claude Code, Codex, and other skill-aware harnesses.

#### Governance

- [Virtual Keys](https://docs.getbifrost.ai/features/governance/virtual-keys.md): Virtual keys are a way to manage access to your AI models.
- [Routing](https://docs.getbifrost.ai/features/governance/routing.md): Direct requests to specific AI models, providers, and keys using Virtual Keys.
- [Budget and Limits](https://docs.getbifrost.ai/features/governance/budget-and-limits.md): Enterprise-grade budget management and cost control with hierarchical budget allocation through virtual keys, teams, and customers.
- [Model Limits](https://docs.getbifrost.ai/features/governance/model-limits.md): Set budget and rate limits at the model level — globally or per virtual key, optionally filtered to a specific provider — from a single unified interface.
- [MCP Tool Filtering](https://docs.getbifrost.ai/features/governance/mcp-tools.md): Control which MCP tools are available for each Virtual Key.
- [Required Headers](https://docs.getbifrost.ai/features/governance/required-headers.md): Enforce mandatory headers on every request through governance.
- [Complexity Router](https://docs.getbifrost.ai/features/governance/complexity-router.md): Automatically classify incoming LLM requests into complexity tiers and route them to the right model.

#### Prompt Repository

- [Playground](https://docs.getbifrost.ai/features/prompt-repository/playground.md): Create, test, and version prompts in an interactive playground.
- [Prompts plugin](https://docs.getbifrost.ai/features/prompt-repository/prompts-plugin.md): Use committed prompt templates from the Prompt Repository on inference requests via HTTP headers or custom resolvers.

#### Plugins

- [Mocker](https://docs.getbifrost.ai/features/plugins/mocker.md): Mock AI provider responses for testing, development, and simulation purposes.
- [JSON Parser](https://docs.getbifrost.ai/features/plugins/jsonparser.md): A simple Bifrost plugin that handles partial JSON chunks in streaming responses by making them valid JSON objects.