claude[bot]
6d8d38ed9d
When beam was already open, calling open() again would unconditionally call terminateKeepingSettings(), destroying all completed rays and fusions. Now open() returns early if beam is already open, preserving the existing state. This prevents data loss when Ctrl+Shift+B (or any other path) triggers beamInvoke() while beam is actively showing responses. Fixes #1079 Co-authored-by: Enrico Ros <enricoros@users.noreply.github.com>
Big-AGI Open 🧠
This is the open-source foundation of Big-AGI, the multi-model AI workspace for experts.
Big-AGI is the multi-model AI workspace for experts: Engineers architecting systems. Founders making decisions. Researchers validating hypotheses. You need to think broader, decide faster, and build with confidence, then you need Big-AGI.
It comes packed with world-class features like Beam, and is praised for its best-in-class AI chat UX. As an independent, non-VC-funded project, Pro subscriptions at $10.99/mo fund development for everyone, including the free and open-source tiers.
What makes Big-AGI different:
Intelligence: with Beam & Merge for multi-model de-hallucination, native search, and bleeding-edge AI models like Opus 4.7, Nano Banana Pro, Kimi K2.6 or GPT 5.4 - Control: with personas, data ownership, requests inspection, unlimited usage with API keys, and no vendor lock-in - and Speed: with a local-first, over-powered, zero-latency, madly optimized web app.
🧠 Intelligence
|
✨ Experience
|
⚡ Performance
|
🔒 Control
|
|
Beam & Merge No context junk Purest AI outputs |
Flow-state interface Highly customizable Best-in-class UX |
Local-first Highly parallel Madly optimized |
No vendor lock-in Your API keys AI Inspector |
Who uses Big-AGI:
Loved by engineers, founders, researchers, self-hosters, and IT departments for its power, reliability, and transparency.
Choose Big-AGI because you don't need another clone or slop - you need an AI tool that scales with you.
Show me a screenshot:
Sure - here is real-world screeengrab as I'm writing this, while running a Beam to extract SVG from an image with Sonnet 4.5, Opus 4.1, GPT 5.1, Gemini 2.5 Pro, Nano Banana, etc.
Get Started
| Tier | Best For | What You Get | Setup |
|---|---|---|---|
| Big-AGI Open (self-host) | IT | First to get new models support. Maximum control and privacy. | 5-30 min |
| big-agi.com Free | Everyone | Full core experience, improved Beam, new Personas, best UX. | 2 min* |
| big-agi.com Pro $10.99/mo | Professionals | Everything + Sync across unlimited devices + 1GB storage | 2 min* |
*: Configuration requires your API keys. Big-AGI does not charge for model usage or limit your access.
Why Pro? As an independent project, Pro subscriptions fund all development. Early subscribers shape the roadmap directly.
Self-host and developers (full control)
- Develop locally or self-host with Docker on your own infrastructure – guide
- Or fork & run on Vercel:
Our Philosophy
We're an independent, non-VC-funded project with a simple belief: AI should elevate you, not replace you.
This is why we built Big-AGI to be local-first, madly optimized to 0-latency, launched multi-model first to defeat hallucinations, designed Beam around the humans in the loop, re-wrote frameworks and abstractions so you are not vendor locked-in, and obsessed over a powerful UI that works, just works.
NOTE: this is a powerful tool - if you need a toy UI or clone, this ain't it.
Release Notes
- Open 2.0.4: Hyper Params Opus 4.6, GPT-5.4, Gemini 3.1 Pro, AWS Bedrock, parameter accuracy, Anthropic continuation/Fast mode
- Open 2.0.3: Red Carpet Kimi K2.5, Gemini 3 Flash, GPT 5.2, Google Drive, Inworld, Novita.ai, Speech/UX improvements
- Open 2.0.2: Speex multi-vendor speech synthesis, Opus 4.5, Gemini 3 Pro, Nano Banana Pro, Grok 4.1, GPT-5.1, Kimi K2 + 280 fixes
What's New in 2.0 · Oct 31, 2025 · Open
- Big-AGI Open is ready and more productive and faster than ever, with:
- Beam 2: multi-modal, program-based, follow-ups, save presets
- Top-notch AI models support including agentic models and reasoning models
- Image Generation and editing with Nano Banana and gpt-image-1
- Web Search with citations for supported models
- UI & Mobile UI overhaul with peeking and side panels
- And all of the Big-AGI 2 changes and more
- Built for the future, madly optimized
Open links: 👉 changelog 👉 installation 👉 roadmap 👉 documentation
For teams and institutions: Need shared prompts, SSO, or managed deployments? Reach out at enrico@big-agi.com. We're actively collecting requirements from research groups and IT departments.
5,000 Commits Milestone
Hit 5k commits last week. That's a lot of code.
Recent work has been intense:
- Chain of thought reasoning across multiple LLMs: OpenAI o3 and o1, DeepSeek R1, Gemini 2.0 Flash Thinking, and more
- Beam is real - ~35% of our users run it daily to compare models
- New AIX framework lets us scale features we couldn't before
- UI is faster than ever. Like, terminal-fast
The new architecture is solid and the speed improvements are real.
What's New in 1.16.1...1.16.13 · (patch releases)
- 1.16.13: Docker fix (#840)
- 1.16.12: Dockerfile update (#840)
- 1.16.11: v1 final release, documentation updates
- 1.16.10: OpenRouter models support
- 1.16.9: Docker Gemini fix, R1 models support
- 1.16.8: OpenAI ChatGPT-4o Latest, o1 models support
- 1.16.7: OpenAI support for GPT-4o 2024-08-06
- 1.16.6: Groq support for Llama 3.1 models
- 1.16.5: GPT-4o Mini support
- 1.16.4: 8192 tokens support for Claude 3.5 Sonnet
- 1.16.3: Anthropic Claude 3.5 Sonnet model support
- 1.16.2: Improve web downloads, as text, markdown, or HTML
- 1.16.2: Proper support for Gemini models
- 1.16.2: Added the latest Mistral model
- 1.16.2: Tokenizer support for gpt-4o
- 1.16.2: Updates to Beam
- 1.16.1: Support for the new OpenAI GPT-4o 2024-05-13 model
What's New in 1.16.0 · May 9, 2024 · Crystal Clear
- Beam core and UX improvements based on user feedback
- Chat cost estimation 💰 (enable it in Labs / hover the token counter)
- Save/load chat files with Ctrl+S / Ctrl+O on desktop
- Major enhancements to the Auto-Diagrams tool
- YouTube Transcriber Persona for chatting with video content, #500
- Improved formula rendering (LaTeX), and dark-mode diagrams, #508, #520
- Models update: Anthropic, Groq, Ollama, OpenAI, OpenRouter, Perplexity
- Code soft-wrap, chat text selection toolbar, 3x faster on Apple silicon, and more #517, 507
3,000 Commits Milestone · April 7, 2024
- 🥇 Today we celebrate commit 3000 in just over one year, and going stronger 🚀
- 📢️ Thanks everyone for your support and words of love for Big-AGI, we are committed to creating the best AI experiences for everyone.
What's New in 1.15.0 · April 1, 2024 · Beam
- ⚠️ Beam: the multi-model AI chat. find better answers, faster - a game-changer for brainstorming, decision-making, and creativity. #443
- Managed Deployments Auto-Configuration: simplify the UI models setup with backend-set models. #436
- Message Starring ⭐: star important messages within chats, to attach them later. #476
- Enhanced the default Persona
- Fixes to Gemini models and SVGs, improvements to UI and icons
- 1.15.1: Support for Gemini Pro 1.5 and OpenAI Turbo models
- Beast release, over 430 commits, 10,000+ lines changed: release notes, and changes v1.14.1...v1.15.0
What's New in 1.14.1 · March 7, 2024 · Modelmorphic
- Anthropic Claude-3 model family support. #443
- New Perplexity and Groq integration (thanks @Penagwin). #407, #427
- LocalAI deep integration, including support for model galleries
- Mistral Large and Google Gemini 1.5 support
- Performance optimizations: runs much faster, saves lots of power, reduces memory usage
- Enhanced UX with auto-sizing charts, refined search and folder functionalities, perfected scaling
- And with more UI improvements, documentation, bug fixes (20 tickets), and developer enhancements
What's New in 1.13.0 · Feb 8, 2024 · Multi + Mind
https://github.com/enricoros/big-AGI/assets/32999/01732528-730e-41dc-adc7-511385686b13
- Side-by-Side Split Windows: multitask with parallel conversations. #208
- Multi-Chat Mode: message everyone, all at once. #388
- Export tables as CSV: big thanks to @aj47. #392
- Adjustable text size: customize density. #399
- Dev2 Persona Technology Preview
- Better looking chats with improved spacing, fonts, and menus
- More: new video player, LM Studio tutorial (thanks @aj47), MongoDB support (thanks @ranfysvalle02), and speedups
What's New in 1.12.0 · Jan 26, 2024 · AGI Hotline
https://github.com/enricoros/big-AGI/assets/32999/95ceb03c-945d-4fdd-9a9f-3317beb54f3f
- Voice Calls: real-time voice call your personas out of the blue or in relation to a chat #354
- Support OpenAI 0125 Models. #364
- Rename or Auto-Rename chats. #222, #360
- More control over Link Sharing #356
- Accessibility to screen readers #358
- Export chats to Markdown #337
- Paste tables from Excel #286
- Ollama model updates and context window detection fixes #309
What's New in 1.11.0 · Jan 16, 2024 · Singularity
https://github.com/enricoros/big-AGI/assets/1590910/a6b8e172-0726-4b03-a5e5-10cfcb110c68
- Find chats: search in titles and content, with frequency ranking. #329
- Commands: command auto-completion (type '/'). #327
- Together AI inference platform support (good speed and newer models). #346
- Persona Creator history, deletion, custom creation, fix llm API timeouts
- Enable adding up to five custom OpenAI-compatible endpoints
- Developer enhancements: new 'Actiles' framework
What's New in 1.10.0 · Jan 6, 2024 · The Year of AGI
- New UI: for both desktop and mobile, sets the stage for future scale. #201
- Conversation Folders: enhanced conversation organization. #321
- LM Studio support and improved token management
- Resizable panes in split-screen conversations.
- Large performance optimizations
- Developer enhancements: new UI framework, updated documentation for proxy settings on browserless/docker
For full details and former releases, check out the archived versions changelog.
👉 Supported Models & Integrations
Delightful UX with latest models exclusive features like Beam for multi-model AI validation.
 |
 |
 |
 |
 |
|---|---|---|---|---|
| Chat Call Beam Draw, ... |
Local & Cloud Open & Closed Cheap & Heavy Google, Mistral, ... |
Attachments Diagrams Multi-Chat Mobile-first UI |
Stored Locally Easy self-Host Local actions Data = Gold |
AI Personas Voice Modes Screen Capture Camera + OCR |
AI Models & Vendors
Configure 100s of AI models from 20+ providers:
| AI models | supported vendors |
|---|---|
| Opensource Servers | LocalAI · Ollama |
| Local Servers | LM Studio (non-open) |
| Multimodal services | Anthropic · AWS Bedrock · Azure · Google Gemini · OpenAI |
| LLM services | Alibaba · DeepSeek · Groq · Mistral · Moonshot · OpenPipe · OpenRouter · Perplexity · Together AI · xAI · Z.ai |
| OpenAI-compatible | Any OpenAI-compatible endpoint - models, pricing, and capabilities are auto-detected |
| Image services | OpenAI · Google Gemini (Nano Banana) · LocalAI |
| Speech services | ElevenLabs · Inworld · OpenAI TTS · LocalAI · Browser (Web Speech API) |
Additional Integrations
| More | integrations |
|---|---|
| Web Browse | Browserless · Puppeteer-based |
| Web Search | Google CSE |
| Observability | Helicone |
🚀 Installation
Self-host with Docker, deploy on Vercel, or develop locally. Full setup guide:
Or use the hosted version at big-agi.com with your API keys.
👋 Community & Contributing
Connect
⭐ Star the repo if Big-AGI is useful to you
Contribute
🤖 AI-Powered Issue Assistance
When you open an issue, our custom AI triage system (powered by Claude Code with Big-AGI architecture documentation) analyzes it, searches the codebase, and provides solutions - typically within 30 minutes. We've trained the system on our modules and subsystems so it handles most issues effectively. Your feedback drives development!
Contributors
License
MIT License · Third-Party Notices
2023-2026 · Enrico Ros × Token Fabrics