An event-driven multi-agent ecosystem. Streams live meeting context, extracts decisions, retrieves episodic semantic memory via Qdrant, and orchestrates automated post-meeting follow-ups.
Rather than relying on a single monolithic LLM, MeetMaxxing coordinates specialized micro-agents over an event-driven orchestrator to turn unstructured meeting discussions into structured team velocity.
Streams and ingests live meeting audio and caption chunks with low latency using persistent WebSocket connections and background processing.
Generates proactive suggestions, speaker cues, and contextual insights mid-meeting while conversations are still underway.
Delivers instant 30-second synthesized recaps of prior discussion so late arrivals catch up seamlessly without disrupting active flow.
Produces executive briefs, discussion highlights, core decisions, and verifiable takeaways automatically after each session.
Indexes episodic 768-dim embeddings in Qdrant vector database for semantic recall across meeting history and organizational context.
Ingests supporting files and docs, answering deep queries grounded directly in meeting context with strict factual citations.
Identifies deliverables with owner assignments and deadlines, generating reminders and calendar events directly from discussion.
Drafts comprehensive follow-up emails highlighting decisions and assigned deliverables, delivering them via Resend with verified attendee routing.
Coordinates specialized micro-agents across meeting lifecycle events with fallback resilience, executing isolated tasks instead of relying on a monolithic prompt.
Built by an agile two-person engineering team for high-velocity execution.
Architected the full-stack system: Next.js 15 Material 3 dashboard, FastAPI services, event-driven agent orchestration, Google ADK integration, and the Chrome Extension.
Engineered the multi-agent AI ecosystem: episodic semantic memory with Qdrant vector search, multi-model LLM fallback (Gemini, Groq, OpenRouter), and structured prompt pipelines.
Experience zero-lag live capture, autonomous synthesis, and proactive workflows with MeetMaxxing.