The SpecLynx Manifesto

Your API specs are standardized. Your tooling should be too.

Instead, every spec brought its own tools, and every tool its own script in CI. None of them agree, not even on what an error looks like.

We maintained Swagger UI and Swagger Editor for years. SpecLynx is the work we knew was needed and never found the time for: one toolchain for every spec.

This is what we believe:

• Same answer everywhere. The editor, the CLI, CI, and your own integration share one implementation. A spec that passes in one place passes in all of them.
• Errors that point somewhere. Line, column, code, and a message you can act on, for every spec, not just OpenAPI.
• Broken is still readable. A spec with a typo is still a spec. Tools should recover and keep going, not stop at the first error.
• Nothing lost. Comments, ordering, and formatting come back as they went in.
• One shared semantic structure. Every spec is read into the same structure, with spec-level semantics layered on top. Nothing gets flattened.
• Your machine, your specs. Local-first. No accounts, no telemetry, nothing uploaded.
• Built in the open. Apache 2.0, free to use, and self-funded. Sponsorship welcome.

We want SpecLynx to follow you from job to job, and to be the layer other tools are built on.

See for yourself:

npx @speclynx/cli validate openapi.yaml

Vladimír Gorej and Francesco Tumanischvili Maintainers, SpecLynx