Top-down file truncation is why coding agents hallucinate exports on large files. If you have an 800-line controller and clip at line 300, the model never sees lines 301 to 800 where the route handlers and export maps live. We wrote evenSpan.js in ...

Dumping 150,000 tokens into an AI prompt every time an agent runs is expensive and sloppy. We benchmarked TokenCap on full-repo tasks and brought payload size from 152,000 tokens down to 12,450 tokens. That is a 12.2x reduction. How it works: Inste...

We Spent Months Teaching AI to Read Our Codebases. We Forgot to Teach It How to Remember. Every developer has had this session: You spend three hours pair-programming with Claude, Cursor, or Codex. Together, you untangle a gnarly race condition in ...

This release is about something I think AI coding tools are missing: a safe floor. For months, I kept seeing the same failure mode. You give an AI agent a refactoring task. It renames a function.-->Breaks 11 files.-->Ships the change. You spend...

This release is focused on one idea: Give AI coding agents better context, while reducing the amount of unnecessary context they need. When working across Claude Code, Cursor, Windsurf, Copilot, Cline, Antigravity, and other coding agents, a lot of...

v1.9.0 proved that local-first repository intelligence can scale, crossing 2,000 downloads and establishing the performance foundation: bounded graph builds, sharded caching, workspace federation, and compression. v2.0.0 builds on that foundation w...

After months of building, iterating, breaking things, and shipping, I’m excited to finally launch TokenCap on Product Hunt. TokenCap helps AI coding agents work with codebases more efficiently by generating graph-ranked, token-budgeted context so...

Today every AI sees Current Repository But developers think in Yesterday ↓ Today ↓ Tomorrow AI has no idea. v1.7 introduces Repository Timeline Intelligence Instead of asking What does the repository look like? TokenCap answers How did i...

One thing I've realized while building with AI is that context isn't a prompt problem anymore. It's an engineering problem. Over the last few weeks I've been using TokenCap on larger projects and found myself repeatedly asking: Why is the AI re...