thunderbird/thunderbolt

Health v2 · maintenance only

Adoption not part of the score

4767 stars · 329 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

Thunderbolt is an open-source, cross-platform AI client from Thunderbird that lets users connect any OpenAI-compatible model or ACP-compatible agent while self-hosting their data. It targets enterprise on-prem deployments and is still early, under active development and security audit.

Use cases

• self-host an open-source AI chat client on my own infrastructure
• use local models with Ollama or llama.cpp instead of vendor APIs
• deploy an on-prem AI assistant for my enterprise
• avoid vendor lock-in for AI chat and agents
• connect MCP tools and ACP agents to a unified AI client
• run one AI client across web, desktop, and mobile

When to choose

• you need data sovereignty or air-gapped AI deployments
• you want a model-agnostic client supporting OpenAI-compatible APIs and ACP agents
• you need MCP integration with your own systems
• you want an open-source alternative to closed frontier AI stacks

When to avoid

• you need a production-ready, fully supported product today
• you want a fully offline-first experience right now
• you need a hosted inference endpoint with no setup
• you want a lightweight consumer app rather than a self-hosted deployment

Facets

application · maturity experimental

chat-interface agent-framework mcp llm-inference self-hosted artificial-intelligence large-language-models self-hosted cross-platform windows self-hosted ai-client data-sovereignty on-prem openai-compatible acp-agents enterprise-ai typescript ai-agents web-server macos linux ios android docker

Member repositories

For agents

markdown · JSON · MCP: product_card(name="thunderbird/thunderbolt")