When to Use the API vs. the Platform
| Use Case | Recommended |
|---|---|
| Build a full app from scratch | Platform (pre.dev web app) |
| Generate specs inside Cursor/Claude Code | MCP Server |
| Integrate spec generation into CI/CD | API |
| Build custom tooling on top of pre.dev | API + SDKs |
| One-off spec for a feature you’ll build manually | MCP Server |
The Architect API
The API gives you programmatic access to pre.dev’s spec generation:POST /fast-spec— Quick specs in ~30-60 seconds (~3-8 credits)POST /deep-spec— Comprehensive specs in ~2-5 minutes (~15-40 credits)GET /spec-status/:id— Check async generation progressGET /list-specs— List your generated specsGET /find-specs— Search specs by contentGET /credits-balance— Check remaining credits
Authorization header.
The MCP Server
The MCP server lets AI coding agents (Cursor, Claude Code, VS Code Copilot, Windsurf) generate pre.dev specs conversationally. Instead of switching to the web app, you can generate architecture specs directly in your IDE. Supported tools:- Cursor (one-click install)
- Claude Code
- VS Code
- Windsurf
- Any MCP-compatible client
SDKs
For deeper integration, use the official SDKs:API vs. Platform: What’s Different?
The API provides spec generation only — it produces the structured specification (milestones, stories, subtasks) as markdown. It does not include:- Autonomous build agents
- Acceptance verification
- PR creation
- Progress tracking
- Knowledge Base integration
Common API/MCP Patterns
Generate spec → Feed to your own agent
Use the API to generate a comprehensive spec, then feed it as context to your own coding agent (Cursor, Claude Code, Aider, etc.).Pre-planning in CI
Automatically generate a spec for every feature branch to validate scope before development begins.Custom dashboards
Use the API to generate and track specs as part of your team’s internal tooling.Getting Your API Key
Get API Key
Visit the playground to generate your API key.

