ManticScore is an AI-powered product development platform built for founders and product teams. From validating a new idea with deep market research to generating GitHub pull requests autonomously, ManticScore compresses weeks of product work into hours.Documentation Index
Fetch the complete documentation index at: https://docs.manticscore.com/llms.txt
Use this file to discover all available pages before exploring further.
Quick Start
Make your first API call and run your first market research in minutes.
Authentication
Learn how to authenticate with the ManticScore API using session tokens.
Market Research
Run AI-powered market research to find competitors, gaps, and opportunities.
API Reference
Browse the complete REST API reference with request and response examples.
What ManticScore does
ManticScore combines several AI pipelines into a single product workspace:- Market Research — Submit a product idea and receive a structured analysis: incumbent competitors, emerging players, key features, signals, and white-space opportunities.
- Feature Deep Research — Drill into specific features across multiple competitors with multi-stage AI analysis and build estimates.
- Build Graphs — Turn research into a recursive build tree of implementable tasks, each with specs, acceptance criteria, and risk ratings.
- AgentForge — Give the AI a coding task and a GitHub repo; it plans, writes code, and opens a pull request autonomously.
- Competitive Signals — Monitor companies and topics for real-time signals — press releases, product launches, job postings — with AI-generated insights.
- AI Chat Agent — Chat with an agent that has full context of your research and build plans, plus access to 1,000+ integrations via Composio.
- Integrations — Ship your build plan directly to Linear, Jira, Slack, Notion, or GitHub with one tap.
- Automations — Configure “when research completes → post to Slack” style event-driven automations.
How the API works
All ManticScore functionality is exposed through a REST API athttps://api.manticscore.com. Long-running operations (research, build graph generation, Forge runs) are asynchronous — you start a job and then stream progress events via NDJSON or WebSocket.
Authenticate
Exchange your Clerk JWT for a session token with
POST /auth/session. Use the token as a Bearer header on all subsequent requests.Start a job
Submit a request to start a research job, build graph, or Forge run. You receive a job ID immediately.
Stream progress
Connect to the events endpoint (
GET /research/{job_id}/events) to receive real-time NDJSON progress updates.