Inspiration
Product and engineering teams struggle to track fast-moving external market shifts and connect them to what developers are actually discussing internally. MarketPulse AI solves this by turning Slack into an active corporate radar.
What it does
MarketPulse AI uses a continuous background ingestion worker to stream live external global headlines into a dedicated alerts channel. When a user triggers an escalation with !trigger-pain-alert, the agent autonomously runs tool lookups to scan internal chat records and instantly compile structured incident summaries and strategic action cards.
How we built it
Built natively using Bolt for JavaScript and the OpenAI Agents SDK, the system leverages a unique dual-layer production fallback architecture. In live enterprise deployments, it establishes connection links with a Slack Model Context Protocol (MCP) HTTP Server for native searching. For portable testing and validation, it falls back to localized structural sandbox tools.
Challenges we overcame
We engineered rigorous execution token loops and response parsing rules to prevent conversational infinite-turn recursion and message timeouts within Slack's event streams.
Built With
- generative-ai
- javascript
- mcp
- model-context-protocol
- newsapi
- node.js
- openai-agents-sdk
- slack-api
- slack-bolt-framework
Log in or sign up for Devpost to join the conversation.