Tessl is the London startup building a spec-centric AI development environment where developers manage intent and AI writes the code.
Tessl’s premise is that in an AI-first world, source code becomes a build artifact — the thing developers own is the specification. Its platform lets teams describe what software should do in living spec files, then generates and maintains the implementation with AI, keeping code in sync as specs evolve. Founded by a repeat London founder, Tessl targets professional teams who need AI-generated code that stays maintainable rather than one-off snippets.
Tessl offers free access for individual developers building on the platform, with paid team and enterprise plans for collaboration and support.
Professional development teams who want AI coding that scales beyond autocomplete — with structure that keeps generated code coherent over time.
Tessl fits teams whose AI-generated code is outgrowing their ability to keep it coherent.
Developers sign up for free access, write their first spec and watch the platform generate and maintain the implementation alongside it.
Teams keep the speed of AI coding while regaining the maintainability that ad-hoc generation tends to destroy.
Developers write specs and generate code within the environment; team features support shared ownership and review of both specs and output.
The approach asks teams to change habits around spec writing, and the platform is still evolving rapidly.
Early users like regenerating whole features from intent instead of chasing broken AI edits.
Tessl’s bet — that specs, not code, become what developers own — matches where AI coding is heading. Teams early enough to build spec discipline now will compound the advantage as generated code volumes explode.
Individuals can use Tessl for free; teams and companies subscribe for collaboration features, governance and support.
Tessl is a development environment; programmatic access to generation capabilities is part of its platform roadmap.
Against code-assistant plugins, Tessl manages whole-codebase change from specs; against agent coding tools, it adds durable structure and team ownership.