E2C is for teams building AI-native systems — not only model prompts. You integrate with an environment where candidates become evidence, evidence becomes state, and state becomes release decisions with a ledger trail.
E2C does not control model generation. It adjudicates whether generated outputs may enter the responsibility state space and trigger allowed transitions. Think: candidate → evidence → adjudication → state → allowed action — not “make the model deterministic.”
Public guides (quickstarts, architecture notes, API references) will be published after review. Until then, see the documentation hub placeholder and contact us for design-partner access.
Expect documentation on:
Claude Code, Codex, and similar tools are bounded implementation workers in an E2C deployment. They can edit and run code; E2C records whether each responsibility slice completed scoped tests, ledger admission, and release gates. Long autonomous chains are not a substitute for mechanical adjudication.
Early engineering partners work toward a unified workspace: project/repo intake, responsibility object discovery, validation and authorization, and release gate outcomes with auditable artifacts. Join the waitlist on Support.