# stdapi.ai > Self-hosted OpenAI, Anthropic, Cohere and Ollama compatible API gateway for Amazon Bedrock. Point your applications at it and reach 100+ models across 20+ providers including Claude, OpenAI GPT, xAI Grok and Amazon Nova, for chat, images, video, speech, transcription, embeddings and moderation — in your own AWS account, at AWS rates with no markup. - [Home](https://stdapi.jgoutin.dev/md/index.md): Run your favorite OpenAI, Anthropic, Cohere and Ollama-compatible apps on Amazon Bedrock. Access 100+ models including Claude, OpenAI GPT, xAI Grok, Amazon Nova for chat, images, video, audio, embeddings, and content moderation — in your own AWS account, at AWS rates with no markup. Open source (AGPL-3.0-or-later). - [Models](https://stdapi.jgoutin.dev/md/models.md): Browse, filter, chart and compare every model stdapi.ai serves on AWS through the OpenAI, Anthropic, Cohere and Ollama APIs - modalities, regional availability, published AWS prices and independent leaderboard scores, side by side. - [Contact](https://stdapi.jgoutin.dev/md/contact.md): How to report a bug or ask a question about stdapi.ai ## Features - [Features](https://stdapi.jgoutin.dev/md/features.md): stdapi.ai features — OpenAI, Anthropic, Cohere and Ollama API compatibility for Amazon Bedrock & AI services. Multi-region access, compliance controls, security, observability, and more. - [Compare](https://stdapi.jgoutin.dev/md/compare.md): How stdapi.ai compares with LiteLLM, Bedrock Access Gateway and Amazon Bedrock Mantle in front of Amazon Bedrock, with sources and verification dates, and the test evidence behind the compatibility claims. ## Get Started - [Overview](https://stdapi.jgoutin.dev/md/getting_started.md): Pick your path to try stdapi.ai - a one-line Docker run for local development, or a two-command Terraform deployment on AWS. - [Deploy on AWS](https://stdapi.jgoutin.dev/md/operations_getting_started.md): Deploy stdapi.ai on AWS with two Terraform commands. Production-grade ECS Fargate deployment with HTTPS, auto-scaling, and optional WAF and monitoring. - [Run Locally with Docker](https://stdapi.jgoutin.dev/md/operations_getting_started_local.md): Run stdapi.ai locally with Docker or Podman for development, testing, and evaluation. Free hardened image with full API compatibility for Amazon Bedrock models. - [Troubleshooting](https://stdapi.jgoutin.dev/md/operations_troubleshooting.md): Fixes for the most common errors encountered when deploying and running stdapi.ai - Terraform failures, 400/401/403/404/429/503 responses, Bedrock throttling, IAM permission errors, S3 bucket errors, VPC connectivity, and more. ## Documentation - [Releases & Roadmap](https://stdapi.jgoutin.dev/md/roadmap.md): stdapi.ai release history. The project is discontinued and v1.20.2, released 2026-11-18, is its final release. - [Trademarks](https://stdapi.jgoutin.dev/md/trademarks.md): Third-party trademarks named or shown in the stdapi.ai documentation, their owners, and the basis on which they are used. ## API Reference - [Overview](https://stdapi.jgoutin.dev/md/api_overview.md): Complete API documentation for the stdapi.ai OpenAI-, Anthropic-, Cohere- and Ollama-compatible gateway. Access Amazon Bedrock models, chat completions, messages, embeddings, image generation, and audio APIs with SDK compatibility. - [API Reference — OpenAPI Spec](https://stdapi.jgoutin.dev/openapi.yml) ## OpenAI Compatible - [Responses](https://stdapi.jgoutin.dev/md/api_openai_responses.md): OpenAI-compatible Responses API for Amazon Bedrock models. Supports streaming, tool calling, structured outputs, and multimodal inputs. - [Conversations](https://stdapi.jgoutin.dev/md/api_openai_conversations.md): Create conversations, add and list items, and attach them to Responses requests with the OpenAI-compatible Conversations API. Server-side multi-turn state with metadata, cursor pagination and per-item deletion. - [Chat Completions](https://stdapi.jgoutin.dev/md/api_openai_chat_completions.md): OpenAI-compatible chat completions API for Amazon Bedrock models including Claude, Nova, Llama. Supports streaming, reasoning modes, prompt caching, and multi-modal inputs. - [Completions](https://stdapi.jgoutin.dev/md/api_openai_completions.md): OpenAI-compatible completions API for Amazon Bedrock models including Claude, Nova, Llama. Simple prompt-to-text generation with the smallest token footprint. - [Files](https://stdapi.jgoutin.dev/md/api_openai_files.md): Upload, manage, and reference files in chat completions using the OpenAI-compatible Files API backed by Amazon S3. Supports documents, images, expiry, cursor pagination, and multipart uploads. - [Vector Stores](https://stdapi.jgoutin.dev/md/api_openai_vector_stores.md): Index files into a managed vector store and search them by meaning with the OpenAI-compatible Vector Stores API. Asynchronous indexing, file batches, attribute filters, expiration policies and per-chunk scoring, backed by your own AWS account — or by an Amazon Bedrock knowledge base you already run. - [Batches](https://stdapi.jgoutin.dev/md/api_openai_batches.md): Run thousands of chat completion or embedding requests asynchronously at a discounted price with an OpenAI-compatible Batch API backed by Amazon Bedrock batch inference. - [Images Generations](https://stdapi.jgoutin.dev/md/api_openai_images_generations.md): Generate images with Amazon Bedrock using Stability AI and Amazon Nova Canvas. OpenAI-compatible text-to-image API with streaming support and multiple output formats. - [Images Edits](https://stdapi.jgoutin.dev/md/api_openai_images_edits.md): Edit and transform images with Amazon Bedrock image models. OpenAI-compatible API for image modifications, inpainting, and transformations. - [Images Variations](https://stdapi.jgoutin.dev/md/api_openai_images_variations.md): Generate variations of existing images with Amazon Bedrock. OpenAI-compatible API for creating alternative versions while maintaining style and composition. - [Videos](https://stdapi.jgoutin.dev/md/api_openai_videos.md): Generate videos from text prompts and reference images with Amazon Bedrock video models using an OpenAI-compatible Videos API. Amazon Nova Reel and Luma Ray 2 through the Sora API shape. - [Audio Speech](https://stdapi.jgoutin.dev/md/api_openai_audio_speech.md): Generate natural-sounding speech from text using Amazon Polly. OpenAI-compatible TTS API with 60+ voices, 30+ languages, and SSML support. - [Audio Transcriptions](https://stdapi.jgoutin.dev/md/api_openai_audio_transcriptions.md): Transcribe audio to text with Amazon Transcribe or Amazon Bedrock audio-capable models. OpenAI-compatible STT API supporting 100+ languages, speaker diarization, and multiple output formats. - [Audio Translations](https://stdapi.jgoutin.dev/md/api_openai_audio_translations.md): Translate audio from any language to English text using Amazon Transcribe + Translate or Amazon Bedrock audio models. OpenAI-compatible API with automatic language detection. - [Realtime](https://stdapi.jgoutin.dev/md/api_openai_realtime.md): Build low-latency voice agents on Amazon Bedrock with an OpenAI-compatible Realtime API — bidirectional WebSocket audio, browser-safe ephemeral client secrets, and G.711 telephony formats. - [Embeddings](https://stdapi.jgoutin.dev/md/api_openai_embeddings.md): Generate vector embeddings with Amazon Bedrock using OpenAI-compatible API. Support for semantic search, RAG, multimodal embeddings with text, images, audio, and documents. - [Moderations](https://stdapi.jgoutin.dev/md/api_openai_moderations.md): Classify text and images with Amazon Bedrock Guardrails, inline guardrail checks, or Amazon Comprehend toxicity detection through an OpenAI-compatible Moderations API. Content safety, policy enforcement, and harm detection for AI applications. - [Models](https://stdapi.jgoutin.dev/md/api_openai_models.md): Discover and list available Amazon Bedrock models through OpenAI-compatible API. Browse 100+ models including Claude, Nova, Llama, and more across all configured regions. - [Organization Usage](https://stdapi.jgoutin.dev/md/api_openai_organization_usage.md): Report token, character, audio and image consumption and the AWS cost of serving it, in time buckets, through an OpenAI-compatible Administration Usage and Costs API backed by Amazon CloudWatch metrics. ## Anthropic Compatible - [Messages](https://stdapi.jgoutin.dev/md/api_anthropic_messages.md): Anthropic-compatible Messages API for Amazon Bedrock models including Claude, Nova, Llama. Supports streaming, extended thinking, tool calling, prompt caching, and multi-modal inputs. - [Message Batches](https://stdapi.jgoutin.dev/md/api_anthropic_batches.md): Run thousands of Messages API requests asynchronously at a discounted price with an Anthropic-compatible Message Batches API backed by Amazon Bedrock batch inference. - [Files](https://stdapi.jgoutin.dev/md/api_anthropic_files.md): Upload, manage, and reference files in Anthropic Messages requests using the Anthropic-compatible Files API backed by Amazon S3. Supports documents, images, selection by ID, page cursors, and bidirectional cursor pagination. - [Models](https://stdapi.jgoutin.dev/md/api_anthropic_models.md): Discover and list available Amazon Bedrock text models through an Anthropic-compatible API. ## Cohere Compatible - [Rerank](https://stdapi.jgoutin.dev/md/api_cohere_rerank.md): Rerank documents by semantic relevance with Amazon Bedrock rerank models using a Cohere-compatible API. Improve RAG and search quality with Amazon Rerank and Cohere Rerank 3.5. - [Embed](https://stdapi.jgoutin.dev/md/api_cohere_embed.md): Generate vector embeddings with Amazon Bedrock embedding models using a Cohere-compatible API. Semantic search, RAG, and multimodal embeddings through the Cohere v2 Embed interface. ## Ollama Compatible - [Generate](https://stdapi.jgoutin.dev/md/api_ollama_generate.md): Generate text for a single prompt with Amazon Bedrock models using the Ollama-compatible /api/generate endpoint. Streaming, images, structured output, and thinking. - [Chat](https://stdapi.jgoutin.dev/md/api_ollama_chat.md): Generate conversational AI responses with Amazon Bedrock models using the Ollama-compatible /api/chat endpoint. Streaming, tool calling, images, structured output, and thinking. - [Embed](https://stdapi.jgoutin.dev/md/api_ollama_embed.md): Generate vector embeddings with Amazon Bedrock embedding models using the Ollama-compatible /api/embed and /api/embeddings endpoints. - [Models](https://stdapi.jgoutin.dev/md/api_ollama_models.md): List, describe and manage Amazon Bedrock models through the Ollama-compatible /api/tags, /api/show, /api/ps, /api/version and model management endpoints. ## Native Extensions - [Search Models](https://stdapi.jgoutin.dev/md/api_search_models.md): Search and filter available Amazon Bedrock models by modality, route, MCP tool, region, streaming support, Batch API support, and legacy status. Designed for AI agents that need to discover the right model before calling other endpoints. - [Model Pricing](https://stdapi.jgoutin.dev/md/api_model_pricing.md): Query the exact AWS unit prices for any model — token rates, service tiers, prompt-cache TTLs, cross-region routing, image specs, and long-context premiums — straight from the AWS Price List catalog used for cost tracking. ## Guides - [Advanced Deployment](https://stdapi.jgoutin.dev/md/operations_deploy_advanced.md): Advanced deployment configurations for stdapi.ai including VPC integration, multi-region production, cost-optimized setups, manual ECS deployment, and Terraform module outputs. - [Authentication & Security](https://stdapi.jgoutin.dev/md/operations_authentication_security.md): Configuration guide for authenticating and securing requests to stdapi.ai — API keys via SSM/Secrets Manager, per-tenant API keys with model and endpoint scopes, tenant-owned AWS credentials via cross-account IAM roles, Amazon Cognito user pool tokens, OAuth 2.0 discovery for AI agents, external authentication via AWS ALB/API Gateway, and encryption in transit. - [Cost Management](https://stdapi.jgoutin.dev/md/operations_cost_management.md): Understand and control what stdapi.ai costs — Marketplace-billed vs AWS-billed models and credit eligibility, gateway infrastructure cost, per-request cost estimation from published AWS prices, and per-user cost attribution. - [Resilience & Failover](https://stdapi.jgoutin.dev/md/operations_resilience.md): Infrastructure-level and application-level resilience for stdapi.ai on AWS — multi-AZ ECS, ALB health checks, Bedrock cross-region inference, S3 durability, and multi-region request routing where each enabled region contributes its own independent Bedrock quota. - [Data Sovereignty & Compliance](https://stdapi.jgoutin.dev/md/operations_compliance.md): How stdapi.ai enforces data residency, sovereignty and content safety on AWS — region-locked Bedrock, S3, KMS encryption, AI service calls, Amazon Bedrock guardrails, and configuration guidance for GDPR, US, and APAC compliance requirements. - [Logging & Monitoring](https://stdapi.jgoutin.dev/md/operations_logging_monitoring.md): Production-grade observability for stdapi.ai with CloudWatch, OpenTelemetry. Track API performance, monitor costs, debug issues, and ensure compliance. - [Licensing](https://stdapi.jgoutin.dev/md/operations_licensing.md): stdapi.ai is free and open source under the AGPL-3.0-or-later license. What the license requires, where the third-party licenses are in the container images, and how to update. ## Configuration Reference - [Overview & Essential Settings](https://stdapi.jgoutin.dev/md/operations_configuration.md): Configuration hub for stdapi.ai - the quick start recipes, the settings that matter ranked by tier, an A-Z index of every environment variable, and the seven configuration reference pages. - [Regions & AWS Clients](https://stdapi.jgoutin.dev/md/operations_configuration_aws.md): Configure the AWS regions stdapi.ai calls, the Bedrock region router and failover, Bedrock Mantle, the shared AWS client settings and the regions of the other AWS AI services. - [Storage](https://stdapi.jgoutin.dev/md/operations_configuration_storage.md): Configure the S3 buckets, prefixes and lifecycle rules stdapi.ai uses for files, images, videos, batches and vector stores, plus the DynamoDB table and the transcription buckets. - [Models & Routing](https://stdapi.jgoutin.dev/md/operations_configuration_models.md): Configure which models stdapi.ai serves and how it resolves them: Marketplace and SageMaker endpoints, model aliases and wildcards, default parameters and service tiers, the model cache and the model list. - [Authentication & Tenants](https://stdapi.jgoutin.dev/md/operations_configuration_authentication.md): Configure how stdapi.ai authenticates clients: static API keys, SSM Parameter Store, Secrets Manager, Amazon Cognito user pool tokens, per-tenant API keys and the OAuth discovery documents agents read. - [HTTP Server & MCP](https://stdapi.jgoutin.dev/md/operations_configuration_server.md): Configure the HTTP surface of stdapi.ai: route prefixes, CORS, trusted hosts, proxy headers, TLS, GZip, SSRF protection, request limits, the MCP server and the API documentation routes. - [Bedrock Features](https://stdapi.jgoutin.dev/md/operations_configuration_bedrock.md): Configure the Amazon Bedrock features stdapi.ai exposes: guardrails, session storage, service tiers and latency optimization, text-to-speech, the Realtime API and inference profile, prompt router and prompt ARNs. - [Observability & Usage](https://stdapi.jgoutin.dev/md/operations_configuration_observability.md): Configure what stdapi.ai reports about itself: log level and request logging, OpenTelemetry traces, CloudWatch metrics, cost tracking and the organization Usage API. - [IAM Permissions](https://stdapi.jgoutin.dev/md/operations_iam_permissions.md): Complete AWS IAM permission reference for stdapi.ai, including required Amazon Bedrock policies, feature-specific statements, and full policy examples. ## Use Cases - [Overview](https://stdapi.jgoutin.dev/md/use_cases.md): Integrate Amazon Bedrock with Claude Code, Open WebUI, n8n, voice agents, RAG pipelines, and other OpenAI, Anthropic, Cohere and Ollama-compatible tools. Step-by-step guides for coding assistants, chat interfaces, workflow automation, and more. - [Open WebUI Integration](https://stdapi.jgoutin.dev/md/use_cases_openwebui.md): Deploy Open WebUI with Amazon Bedrock using stdapi.ai. Complete setup guide for private ChatGPT alternative with RAG, voice, images, and multi-modal AI capabilities. - [LobeHub Integration](https://stdapi.jgoutin.dev/md/use_cases_lobehub.md): Deploy LobeHub with Amazon Bedrock using stdapi.ai. Setup guide for a private AI chat platform with vision, image generation, and knowledge-base embeddings through one OpenAI-compatible connection. - [n8n Integration](https://stdapi.jgoutin.dev/md/use_cases_n8n.md): Connect n8n workflows to Amazon Bedrock using stdapi.ai. Automate business processes with AI using hundreds of integrations, no-code interface, and OpenAI-compatible or Anthropic-compatible nodes. - [AI Coding Assistants](https://stdapi.jgoutin.dev/md/use_cases_coding_assistants.md): Connect Claude Code, Cline, OpenCode, and other AI coding assistants to Amazon Bedrock via stdapi.ai. Use Claude, Kimi thinking, and Qwen3 Coder in VS Code, JetBrains IDEs, Zed, and the terminal. - [Python Client Libraries](https://stdapi.jgoutin.dev/md/use_cases_python_libraries.md): Build Python applications and agents on Amazon Bedrock with stdapi.ai. Configure LangChain (ChatOpenAI, OpenAIEmbeddings, ChatAnthropic), pydantic-ai, the OpenAI Agents SDK and the official Ollama client against stdapi.ai's OpenAI-, Anthropic- and Ollama-compatible endpoints. - [Autonomous Agent CLIs](https://stdapi.jgoutin.dev/md/use_cases_autonomous_agents.md): Run autonomous agent CLIs like Hermes and OpenClaw against Amazon Bedrock models through stdapi.ai. Configure wire-format transports and Anthropic prompt-caching breakpoints. - [RAG Pipelines](https://stdapi.jgoutin.dev/md/use_cases_rag.md): Build retrieval-augmented generation on Amazon Bedrock with stdapi.ai. Use the managed Vector Stores API, or wire Haystack and any OpenAI- and Cohere-compatible framework to embeddings, two-stage reranking, and generation through one endpoint. - [RAGFlow Integration](https://stdapi.jgoutin.dev/md/use_cases_ragflow.md): Deploy RAGFlow with Amazon Bedrock using stdapi.ai. A self-hosted RAG platform on ECS Fargate with managed OpenSearch, Aurora and Valkey, and chat, embedding and rerank models preconfigured. - [Home Assistant Voice](https://stdapi.jgoutin.dev/md/use_cases_home_assistant.md): Connect Home Assistant Assist voice pipelines to Amazon Bedrock using stdapi.ai and the wyoming-openai proxy. Speech-to-text and streaming text-to-speech for local voice assistants, backed by Amazon Transcribe and Amazon Polly.