Cut the noise from your AI context.
Paste a spec, a document or code. See what can go and how many tokens you save before it reaches a model. Nothing is uploaded.
Token counts are estimates (about 4 characters per token). Rules are pattern-based, not semantic, so review the output before using it.
Built for spec-driven development.
Long specs, steering files and generated code fill an agent’s context with text it does not need. Trim them before you hand them over.
Private by design
The tool runs entirely in your browser. Your text is never sent to a server.
Transparent
The Changes view shows every removed line and the rule that removed it.
Works on plain text
Markdown specs, PDFs exported to text, source files and logs.
What is live, and what is next.
Browser context pruner
Rule-based pruning with token estimates, change view, copy and download.
PDF to spec
OCR and structural parsing into Markdown specs and JSON schemas.
SDKs and API
Kotlin and TypeScript libraries plus a REST API for pipelines.
Subscriptions open soon.
Paid plans for the full engine (PDF to spec, SDKs and API) are on the way. The browser pruner stays free. Join the list to hear first.
Built by an engineer tired of wasted context.
OpenRethink started in 2026 from a practical frustration: AI coding assistants spend compute re-reading unstructured files and still lose track of the specification.
Mostafa B.
FounderMobile and cloud developer focused on Kotlin and Android architecture, document processing and AI developer workflows.
Get early access to the full engine.
Subscriptions open soon. Join the list and tell us what you are building. We reply personally.