About Entalpa
Entalpa is a spec-driven development platform for AI coding agents: a persistent, MCP-native specification layer (stakeholders → business needs → user stories → requirements → design specs) with atomic locking and open questions surfaced before the build, for people building with Claude Code, Cursor and Codex.
What it is
A specification that lives in a store rather than a markdown file. It persists across agent sessions, every requirement traces back to the stakeholder need it serves, and a locked requirement is excluded from what the model is allowed to regenerate. Agents read it over MCP, so they build what was meant without re-deriving the project from the codebase each time. The pipeline is on how it works; the setup for each client is under connect your agent.
It is not a replacement for a coding agent, and not another spec workflow: it composes with GitHub Spec Kit, OpenSpec, Kiro specs and CLAUDE.md, and the comparison pages say where each of those wins.
Who it is for
Solo founders and small teams shipping real products with AI agents, and product people who need a spec their engineers’ agents will actually read. The web app at app.entalpa.com has a free tier with no credit card; reading and writing a project over MCP never spends credits; specs export to ReqIF, so there is no lock-in.
How we write
Every guide and every failure-mode page carries a provenance note: what developers said in public about their own projects, an incident in our own repository, a published survey, or a measurement we ran with the method printed on the page. Where a page would need a number we do not have, it says so instead of implying we ran it.
Company
Entalpa Corp., a Delaware corporation (File No. 10580292).
228 Park Ave S, PMB 85451, New York, NY 10003, United States.
Email: contact@entalpa.com
Code: github.com/entalpa (the MCP server and agent skills are open source)
For machines: /llms.txt, /ai/summary.json