Inspiration
Coaches, consultants, and enterprise operators spend 3 to 5 hours a week managing post-session administrative overhead — writing summaries, tracking decisions, and assigning tasks. We built SessionSync Agent to act as an AI Operations Manager that converts raw transcripts into structured corporate assets in under 10 seconds.
What it does
- Ingestion: paste any raw meeting or coaching transcript.
- Agentic Processing: live extraction sequence parses text and enforces a strict output schema.
- Deterministic Output: Executive Summary, Action Items (Who/What/When), and Key Decisions — zero mocked data.
How we built it
Frontend built rapidly with Bolt. AI Engine powered by Anthropic's Claude API (claude-sonnet-4-5) with strict JSON schema enforcement. API calls run through a server-side Edge Function — the key never touches the browser.
Challenges we ran into
Building a fully functional tool in under two hours meant avoiding brittle infrastructure like live scraping or multi-format parsing. We focused on prompt engineering and schema enforcement for a demo that runs live with zero crash risk.
Accomplishments that we're proud of
A real, non-mocked AI agent built in record time — processes live transcripts dynamically with reliable structured output.
What we learned
Strict schema constraints combined with LLM reasoning produce reliable business tools without heavy technical overhead.
What's next for SessionSync Agent
Direct integrations with Asana, ClickUp, Slack; CRM sync with HubSpot/Salesforce; voice-first ingestion.
Built With
- ai
- ai-agents
- anthropic
- bolt
- claude
- javascript
- react

Log in or sign up for Devpost to join the conversation.