BYOAIK (Bring Your Own AI Key) is a directory of AI tools that run on your own API key.

Tools that accept your Custom / OpenAI-compatible API key (226)

Where your key goes

US compelled disclosure: Ownership not established The operating entity has not been established yet.
Browser apps: Depends on the endpoint you choose This entry stands for any OpenAI-compatible endpoint you point a tool at, so there is no single address to probe. The answers depend entirely on whose endpoint you choose.

Tools that support Custom / OpenAI-compatible

torchtune: PyTorch-native fine-tuning recipes with no framework layer between you and the training loop.
OpenLLM: One command to run an open model as your own OpenAI-compatible endpoint.
Higgs Audio: Expressive speech and multi-speaker dialogue generation, self-hosted.
Agent Desk: Self-hosted AI customer support desk with human handoff, on OpenAI-compatible model access you configure.
OpenDocuments: Self-hosted RAG search over your company's documents, with a local-LLM path that keeps data on your network.
DeerFlow: ByteDance's open-source super-agent harness for deep research, coding, and content, with bring-your-own-key LLMs.
Agentic Ledger ✓: Local-first flight recorder for AI agents: every call, tool use, and dollar captured by a transparent proxy on your own key.
PrivateGPT: Privacy-first RAG app to ingest and query your documents using local or remote LLMs, fully offline-capable.
ComfyUI: Node-based diffusion studio that runs image and video models on your own hardware and your own weights: no cloud key at all.
Stability Matrix: One installer and model manager for ComfyUI, SD.Next, InvokeAI and the rest, all running locally.
llama.cpp: The C/C++ inference engine underneath most local-LLM apps: run GGUF models on your own hardware, no key required.
Unsloth: Fine-tune open models on your own GPU, with the training data and the weights never leaving it.
LlamaFactory: Fine-tune a hundred-plus open models from a local web UI, no key and no cloud step.
SD.Next: All-in-one local studio for image and video generation across dozens of model architectures.
Voicebox: Local voice studio for cloning and dictation, positioned against ElevenLabs, with nothing sent anywhere.
TRL: Hugging Face's reference library for SFT, DPO and GRPO, run wherever you run Python.
ACE-Step: Full songs generated on your own GPU, with the weights on your disk and no service in the loop.
MLX LM: Run and fine-tune language models on Apple silicon, entirely on the laptop in front of you.
Piper: Fast neural speech on a Raspberry Pi, the TTS engine behind much of the offline voice-assistant world.
Chatterbox: Zero-shot voice cloning and expressive TTS you run yourself, from a commercial vendor, under MIT.
KoboldCpp: Single-file GGUF runner with a bundled web UI: everything runs on your own machine with no key or account.
Applio: Voice conversion toolkit built on RVC, with training and inference on your own machine.
PEFT: Hugging Face's LoRA and adapter library, the piece most local fine-tuning is actually built on.
verl: Reinforcement learning for language models at scale, on clusters you control.
ms-swift: ModelScope's training framework covering several hundred models, run on your own machines.
F5-TTS: Fast local voice cloning from a short sample, with the sample never leaving your machine.
YuE: Full-song generation with vocals and backing, running on your own GPU.
OpenRLHF: Readable RLHF implementation built on Ray and vLLM, for aligning models on your own data.
text-generation-webui: Desktop app for local models with a full web UI: runs offline with zero telemetry and no API key.
Hugging Face Chat UI: Open-source SvelteKit chat interface powering HuggingChat, self-hostable against any OpenAI-compatible endpoint with your own API key.
Kotaemon: Self-hosted chat-with-your-documents RAG interface with citation highlighting, running on your own API key or a local model.
OpenAI Agents SDK: OpenAI's own agent framework, which also drives any OpenAI-compatible endpoint you point it at.
LocalAI: Self-hosted drop-in replacement for the OpenAI API covering text, vision, audio and images: your hardware, no key.
vLLM: High-throughput self-hosted serving engine for open models on your own GPUs, behind an OpenAI-compatible API you own.
Moshi: Full-duplex speech-to-speech conversation, running on your machine instead of a vendor's.
Stable Audio Tools: Stability's training and inference code for generative audio, to run and fine-tune yourself.
Krita AI Diffusion: AI image generation inside Krita, powered by your own local or remote ComfyUI server
LlamaCoder: Open v0 alternative that builds a React app from a prompt, on a key you provide.
OpenWhispr: Privacy-first voice-to-text dictation with local Whisper/Parakeet and bring-your-own-key cloud AI.
MetaGPT: Open-source multi-agent software company that turns a one-line requirement into code, using your own LLM key.
PaperQA2: Citation-backed question answering over your own library of scientific PDFs, on whichever provider key you export.
DeepEval: Pytest-style evaluation for LLM apps, with judge models running on your own key or locally.
Crawl4AI: Crawler built for feeding LLMs, running locally with no service between you and the web.
smolagents: Hugging Face's deliberately small agent library, model-agnostic and local-friendly.
Draw Things: Image generation that runs offline on an iPhone or a Mac, not in someone else's datacentre.
Axolotl: YAML-configured fine-tuning built for multi-GPU runs you can reproduce and audit.
TensorRT-LLM: NVIDIA's inference engine for squeezing the most out of GPUs you already own.
BentoML: Packages a model into a service you deploy yourself, wherever you want it to run.
LMDeploy: Compression, quantisation and serving toolkit from the InternLM team, run on your own GPUs.
Talon: Self-hosted runtime for one persistent AI agent that follows you across Telegram, Discord, Teams, terminal and desktop.
Stagehand: Browser automation that mixes code and natural language, on an API key you pass in.
manga-image-translator: One-click OCR and translation of text baked into manga, comics, and images, with your own LLM/MT provider keys.
OpenClaw: Personal assistant running on your own hardware, reaching you through the chat apps you already use.
KServe: Kubernetes-native model serving, the piece that makes self-hosting survive production.
Relaticle: Self-hosted AI-native CRM where your team supplies its own Claude, Gemini or OpenAI-compatible key.
Napkins: Turns a wireframe or screenshot into a working app, running on your own key.
Khoj: Open-source, self-hostable AI second brain that chats with your documents and the web.
Hermes Agent: Terminal agent from Nous Research that runs on any provider, or seven different backends.
LightRAG: Graph-based RAG that will run end to end against a local Ollama, with no cloud key.
Graphiti: Temporal knowledge graph for agent memory, where facts carry the time they were true.
Xinference: Self-hosted model server for LLMs, embeddings and speech behind one OpenAI-compatible API.
LearnHouse: Self-hosted learning platform with AI woven through courses, assignments and auto-grading.
SGLang: High-throughput serving engine for open models on your own GPUs, behind an OpenAI-compatible endpoint.
Open TutorAI: Self-hosted tutoring platform that connects to Ollama or any OpenAI-compatible endpoint you point it at.
Buzz: Cross-platform desktop app to transcribe and translate audio offline with Whisper, or via your own OpenAI-compatible API key.
AutoGen: Microsoft's open-source multi-agent framework where every model client takes your own API key or environment variable.
Marker: PDF to markdown that runs its own vision models locally, with LLM boosting on your key.
ChainForge: Visual prompt-engineering environment for comparing prompts across models, with keys you enter or load locally.
chatgpt-shell: Emacs shell for OpenAI, Anthropic, Ollama and more, with keys read at runtime from your own function.
Morphic: Open-source AI answer engine with a generative UI, bring your own provider key
bilingual_book_maker: CLI that translates EPUB/TXT/SRT/PDF books into bilingual editions using your own LLM key.
ShellGPT (sgpt): Natural-language CLI that turns prompts into shell commands, code, and answers via your own LLM key.
NextChat: Light and fast, self-hostable AI chat UI where you bring your own provider keys.
AppFlowy AI: Open-source Notion alternative with AI you control via local Ollama or your own keys
Browser Use: Point any model at a real browser and let it click, on a key and an endpoint you choose.
cognee: Memory for agents built as an ECL pipeline, with an Ollama path that needs no cloud key.
InvokeAI: Open-source Stable Diffusion creative studio with BYOK cloud image models
SiYuan: Privacy-first, self-hosted block-based knowledge base with built-in AI writing and Q&A via your own API key.
Bifrost: Self-hosted AI gateway with governance and MCP support, sitting in front of the keys you own.
OpenMAIC: Multi-agent AI classroom from Tsinghua that turns a topic into a course, on providers you configure.
Skyvern: Automate browser-based workflows with LLMs and computer vision instead of brittle selectors.
OWL (CAMEL-AI): Open-source multi-agent framework for real-world task automation, built on CAMEL-AI.
Agent Zero: Self-hosted, Docker-sandboxed autonomous agent framework that writes and runs its own code using your provider keys.
marimo: Reactive Python notebook stored as plain files, with AI assistance on a key you configure.
txtai: Embeddings database that runs models locally by default, with cloud providers optional.
GraphRAG: Microsoft's graph RAG pipeline, driven by a key you place in its own config.
gptme: Open-source terminal AI agent that writes code, runs shell/Python, and browses the web with your own provider keys.
Local Deep Research: Local-first agentic deep-research assistant that produces cited reports across web, academic, and private sources.
Meetily: Privacy-first, self-hosted AI meeting notetaker with local transcription and BYOK summaries.
Mattermost Agents: Bring-your-own-LLM copilot for self-hosted Mattermost, including fully local models.
Jupyter AI: Generative AI inside JupyterLab, the notebook most teaching and research actually happens in.
WrenAI: Governed text-to-SQL over your own databases, self-hostable and driven by a key you hold.
SWE-agent: Princeton NLP's terminal coding agent that resolves GitHub issues using whichever LLM key you supply.
PydanticAI: Type-checked agent framework where the model is a string you swap, not a dependency you inherit.
DocsGPT: Self-hosted document search, chat and agents over your own files: bring your own OpenAI, Anthropic, Gemini or local key.
LangGraph: Graph-based agent runtime for durable, stateful workflows on providers you configure.
Agno: High-performance agent framework, with a telemetry default it documents and one env var to disable.
Local GPT (Obsidian): Obsidian plugin for AI note actions that defaults to local Ollama models, or any cloud key you configure yourself.
Ellama: Emacs AI assistant that defaults to a local Ollama model and takes your key for anything hosted.
Microsoft Prompt Flow: Microsoft's prompt engineering and evaluation toolkit, where your provider key is registered as a local connection.
vim-ai: Vim and Neovim completion and editing driven by a key you keep in a file you control.
Koharu: Rust-built ML manga translator with optional bring-your-own-key hosted LLM backends.
Readest: Cross-platform open-source EPUB/PDF reader with built-in translation and BYOK AI reading tools.
OpenTutor: Block-based adaptive learning workspace that turns your own course material into notes, quizzes and flashcards on your own key.
Vane: Open-source, self-hostable AI search engine that answers with citations using SearXNG plus your own LLM provider.
Opik: Self-hostable tracing and evaluation for LLM apps, so the traces stay with you.
promptfoo: CLI and library for testing, comparing and red-teaming prompts across providers on your own keys.
mem0: Memory layer for agents that you can run as a library or self-host, on your own model key.
Activepieces: Open-source, self-hostable Zapier alternative with native AI steps and an MCP-based agent toolkit.
Auto-Analyst: Self-hosted data science agent that cleans, analyses and plots on your own provider key.
parrot.nvim: Provider-agnostic Neovim chat and rewriting, with keys read from commands rather than config files.
minuet-ai.nvim: Neovim completion across a dozen providers, including a local Ollama with no key at all.
llm.nvim: Neovim ghost-text completion pointed at whatever endpoint you choose, including your own.
Agent Browser: Vercel Labs' Rust CLI giving an agent direct browser control from the terminal.
MarkItDown: Microsoft's file-to-markdown converter, with image description and OCR on a client you provide.
BlogPilot: Self-hostable SEO content studio covering keyword research, audits and topic authority on one AI key you supply.
Open Interpreter: Natural-language coding agent that runs code locally in your terminal using your own model key.
Qwen Code: Open-source terminal coding agent (Gemini-CLI fork) that runs on your own API key for any OpenAI/Anthropic/Gemini-compatible endpoint.
SurfSense: Open-source, privacy-focused NotebookLM / Perplexity / Glean alternative that connects any LLM to your docs and external sources.
Anki AI Utils: Improves the Anki flashcards you keep failing with AI explanations, mnemonics and images, on your own key.
ChatterUI: Open-source Android LLM frontend: run GGUF models on-device or connect with your own OpenAI, Claude or OpenRouter key.
Anarlog: Open-source, local-first AI meeting notetaker for macOS with bring-your-own-key and local model support.
Whispering: Open-source, local-first dictation and transcription with your own API keys.
Forge: Open-source terminal AI pair programmer with BYOK access to 300+ models
Helicone: Self-hostable observability proxy for LLM calls, which means it sees the prompts it logs.
TryPost: Self-hosted social scheduling with an AI copilot running on keys you put in your own .env.
GPT Researcher: Autonomous agent that runs deep multi-source web research and writes cited reports using your own LLM provider.
BrowserOS: Open-source agentic browser that runs AI agents to automate web tasks with your own API keys.
Copilot for Xcode: Community Xcode extension adding chat and suggestions, with the ChatGPT path running on your own OpenAI key.
Open Generative AI: Self-hosted image and video studio where the key sits in your browser and local models need no key at all.
LiteLLM: One OpenAI-compatible endpoint in front of a hundred providers, running on keys you hold.
LLM: CLI tool and Python library for accessing many LLMs with your own API keys
Text Generator: Template-driven AI text generation inside your Obsidian notes, using your own provider API keys.
Whisper (Obsidian): Obsidian speech-to-text plugin that transcribes recorded or uploaded audio via any Whisper-compatible API using your own key.
Supermemory: Memory API you can self-host, taking your own provider key rather than issuing one.
Dyad: Local open-source AI app builder: an alternative to hosted builders that runs on your machine and your own keys.
Warp: Agentic terminal / development environment with an AI agent built into the shell, supporting your own provider API keys.
Open Notebook: Open-source, privacy-first NotebookLM alternative that chats with your sources and generates multi-speaker AI podcasts.
Inbox Zero: Open-source AI email assistant that automates triage, replies, and inbox cleanup.
Omni: Self-hosted company AI agent that works across your tools on providers you nominate.
OpenHands: Open-source autonomous AI software-engineering agent that runs on any LLM you supply.
Chatwoot: Self-hosted support and CRM suite whose AI copilot runs on the OpenAI-compatible key you enter yourself.
Karakeep: Self-hostable bookmark-everything app with AI tagging and full-text search
avante.nvim: Use your Neovim like Cursor AI IDE, with diff-reviewed AI code edits.
DSPy: Stanford framework for programming prompts rather than hand-tuning them, against any model on your own key.
VoiceInk: Open-source macOS dictation app with local whisper.cpp and bring-your-own-key cloud STT and AI cleanup.
Onyx: Open-source enterprise search and AI assistant that works with every LLM and connects to your knowledge sources.
stagewise: Frontend coding agent that reads the live DOM, running on the model key you configure.
Jan: Open-source local-first desktop AI assistant that runs models offline or via your own cloud provider keys.
gptel: Simple, extensible LLM client for Emacs that works in any buffer
Langflow: Open-source visual builder for AI agents and RAG workflows on a node canvas.
Langfuse: Self-hosted LLM engineering platform: prompt management, playground and evals all run against a key you bring.
Chatbox: Cross-platform AI client for ChatGPT, Claude, Gemini, and local models with your own keys.
opencode.nvim: Drives OpenCode from inside Neovim, inheriting whichever provider OpenCode is configured with.
assistant-ui: TypeScript/React component library for building production AI chat UIs wired to your own provider keys and backend.
BeeAI Framework: Open-source framework for building production-ready multi-agent systems in Python and TypeScript with a unified, bring-your-own-key LLM backend.
Zed: High-performance Rust code editor with a built-in AI agent that runs on your own provider API keys.
ChatALL: Desktop app that queries ChatGPT, Claude, Gemini and more side by side, each using your own API key.
Copilot for Obsidian: Open-source AI assistant plugin for Obsidian with chat, vault Q&A, and agentic workflows over your notes.
PyGPT: Desktop assistant for Windows, macOS and Linux that states plainly it runs on your own API key.
Crush: Glamourous terminal-native agentic coding agent that works with any LLM via your own API key.
Headroom: Local proxy that compresses context before it reaches the model, holding your key on your own machine.
Cherry Studio: All-in-one desktop AI client where you plug in your own provider keys for chat, agents, and 300+ assistants.
Open WebUI: Feature-rich, self-hosted web UI for LLMs with full bring-your-own-key support across OpenAI, Ollama, and any OpenAI-compatible provider.
LlamaIndex: Leading open-source data and agent framework for building research agents; you supply every provider's API key.
OpenAI Codex CLI: Open-source terminal coding agent that edits files and runs commands in a sandbox.
Windmill: Open-source developer platform that turns scripts into workflows, APIs, and UIs, with bring-your-own-key AI codegen.
Subtitle Edit: Mature open-source subtitle editor with Whisper transcription and BYOK LLM auto-translation
Text Generator (Obsidian): Bring-your-own-key AI text generation inside Obsidian via templated prompts.
big-AGI: Multi-model AI workspace with Beam multi-model chat, personas, image and voice, on your own keys
Open Cowork: One-click desktop agent app for Windows and macOS, keys supplied by you.
ProxyAI: Open-source JetBrains coding assistant, formerly CodeGPT, that runs on the provider key you paste into settings.
org-ai: Emacs org-mode package that calls OpenAI, Claude, Perplexity or a local model using the key you set yourself.
Scira: Open-source agentic AI search engine that plans, retrieves, and cites sources inline.
InSidebar AI: Chrome and Edge side panel that keeps your keys in browser storage and calls providers directly.
Goose: Block's open-source, extensible AI agent for the CLI and desktop that installs, runs, edits, and tests code with any LLM you bring.
Airi: Self-hosted, you-owned AI companion with a VRM/Live2D avatar and real-time voice chat.
CodeCompanion.nvim: AI coding assistant for Neovim, Vim style
Deep Research: Browser-based deep-research app that generates in-depth reports using any LLM, with keys stored locally in your browser.
AI Engine (WordPress): WordPress's most-installed AI plugin: every chatbot and content feature runs on the key you paste in.
Summarize: Point at any URL, YouTube, podcast, PDF, or file and get the gist, via Chrome/Firefox sidebar or CLI using your own model API key.
KISS Translator: Minimalist open-source bilingual webpage translation extension with bring-your-own-key AI support.
Tabby: Self-hosted, open-source AI coding assistant (Copilot alternative) you point at the models and providers you choose.
Fabric: Open-source CLI framework for augmenting humans with AI via reusable, crowdsourced prompt "patterns."
Letta: Open-source platform for building stateful AI agents with long-term memory
Pi: Minimal open-source AI agent toolkit and coding-agent CLI with a unified multi-provider LLM API, BYOK.
CrewAI: Open-source Python framework for orchestrating collaborative, role-playing autonomous AI agents.
bolt.diy: Open-source in-browser full-stack app builder that runs on any LLM you wire up with your own key.
Dify: Open-source platform for building agentic AI workflows and RAG apps with your own model keys.
Arize Phoenix: Self-hosted LLM tracing and evaluation, under a source-available licence rather than an open one.
Haystack: Open-source Python framework for building production RAG pipelines and agent workflows with any LLM provider.
Screenpipe: Records your screen and audio continuously on your own machine, and can run the AI locally too.
Nanobrowser: Open-source Chrome extension running multi-agent AI web automation with your own LLM keys.
WebBrain: Chrome and Firefox browsing agent that can point at any OpenAI-compatible server, including yours.
SillyTavern: Open-source self-hosted LLM frontend for power users and roleplay, BYOK to many backends.
Mastra: Open-source TypeScript framework for building AI agents and graph-based workflows.
RAGFlow: Open-source RAG engine with deep document understanding for grounded, cited document Q&A.
Sim: Open-source visual builder for AI agent workflows with full bring-your-own-key support.
OpenAgent: Self-hostable personal assistant pairing your own RAG knowledge base with thirty-plus model providers.
LibreChat: Open-source, self-hosted ChatGPT alternative unifying many AI providers with agents, MCP, and RAG.
OpenCode: Open-source, provider-agnostic AI coding agent built for the terminal.
AnythingLLM: All-in-one local-first AI app for document chat (RAG) and agents, with bring-your-own-key support for nearly every LLM provider.
Cline: Open-source autonomous coding agent for VS Code, JetBrains, and CLI with bring-your-own-key model support.
Pot (pot-desktop): Cross-platform desktop app for selection translation and OCR, powered by your own provider keys.
GPT Pilot: Builds whole applications step by step with you reviewing, under a licence that becomes MIT later.
ChromeClaw: OpenClaw's agent ideas in a Chrome extension: load it, add your key, start.
Aider: AI pair programming in your terminal, editing code in your local git repo.
Twenty CRM: Open-source, AI-first CRM (Salesforce alternative) with bring-your-own-key AI agents.
Kilo Code: Open-source AI coding agent for VS Code, JetBrains, and CLI with zero-markup BYOK
n8n: Fair-code workflow automation with native AI Agent and LangChain nodes.
Witsy: BYOK desktop AI assistant and universal MCP client spanning chat, image, video, voice, and search
ChatGPT Box: Deep AI integration in your browser with sidebar chat, selection tools, and page summarization.
aichat: All-in-one LLM CLI with Shell Assistant, Chat-REPL, RAG, and agents across 20+ providers using your own keys.
Continue: Open-source VS Code and JetBrains AI coding assistant where every model call uses your own provider API key.
ChatHub: Several chatbots side by side in one extension, usable through your own API keys.
LangChain: The most widely used open-source LLM app and agent framework; every chat model integration takes your own API key.
NextAI Translator: Open-source browser-extension and desktop translator (successor to OpenAI Translator) that translates, polishes, and summarizes text using your own provider API keys.
TypingMind: One-time-purchase power-user chat UI that wraps your own LLM provider keys.
LobeChat: Polished open-source AI chat UI/PWA with 40+ provider support, plugins, and multi-modal features.
Flowise: Open-source drag-and-drop builder for LLM apps and AI agents.
Read Frog: Open-source AI browser extension for immersive bilingual translation and language learning.
Page Assist: Open-source browser extension giving a sidebar and web UI for local and cloud AI models via your own keys.
AI Actions for Shortcuts: Bring-your-own-key AI actions for Apple Shortcuts on iOS, macOS, and visionOS.
GitHub Copilot (BYOK): Run Copilot Chat and agents on your own provider API keys
MacWhisper: Local Whisper transcription for Mac with bring-your-own-key cloud STT and LLM cleanup
Monet: Consent layer that takes your provider key once and gives apps a revocable token instead of the raw secret.