Everything you need to give your local AI persistent memory.
ChatGPT, Claude, Gemini, or any JSON conversation export. Drag and drop. Auto-detected. Facts extracted automatically from thousands of conversations.
Every conversation through the proxy is scanned for new facts. Quality scored against existing knowledge. Never stores fiction as fact.
Semantic search across your entire memory graph. Only relevant facts injected into context. Not your entire history dumped into the window.
Ollama, OpenAI, Anthropic, Groq. Proxy works with any OpenAI-compatible provider. Switch anytime. Memory stays consistent.
Native window with 8 tabs: Facts, Review, Timeline, Insights, Status, Settings, Import, Mesh. No terminal required.
Local-first. API keys encrypted at rest. Binds to 127.0.0.1 by default. Nothing phones home. No telemetry.
Velqua detects which agent is calling. Blackreach, coder, planner — each gets its own identity and private memory.
Agents write findings to a shared pool. Other agents read it. Coordination without a framework. No SDK required.
Live WebSocket view of all active agents, their tasks, shared memory writes. Zero configuration.