Repo · Context & memory
Winnow
Filter tool output before it enters context.
Open github.com ↗From the repository
A calibrated context sieve for Claude Code: every tool result is judged by a System One model before it enters context.
claude-codeclaude-code-plugincontext-managementjevllm-agentstypesafe
How builders describe it
A calibrated context sieve for Claude Code: every tool result is judged by a System One model before it enters context.
Calibrated context sieve for Claude Code: every large tool result is judged by TypeSafe Jev (or a System One adapter) before it enters context; confident-no blocks become recallable stubs. Distinct from any Chrome read/skim “Winnow” tools.
A context sieve for Claude Code: Jev judges each tool result before it lands.
The decision Jev makes
Keep or discard a tool result, or answer whether a stopping condition is met.
Where it fits
Agents drown in their own context. These projects use Jev to decide what to keep, what to drop and when a loop is actually finished. All 66 context & memory projects →
Related projects
Context & memoryjev-compactionContext compaction library (jevctx) where Jev only scores—never rewrites—so retained lines stay original, elided text is recoverable byte-for-byte,…GitHub · ★ 4 · Waxmell114514
Context & memoryjev-skillsClaude Code / Codex plugin: TypeSafe Jev decides which installed skills enter context each turn—always-on skill-list cost drops to zero tokens.GitHub · ★ 2 · eran-broder
Context & memoryopencode-context-prunerContinuous verbatim context pruning for OpenCode, powered by TypeSafe Jev.GitHub · ★ 2 · hoshinodis
Context & memoryjev-filterSemantic filtering for AI tool results.GitHub · ★ 1 · apixly-ai
Context & memoryA memory gate for Claude CodeOne question per message: is this worth remembering?X post · Jetwani Avinash
Context & memoryBeacon fixes a coding agent’s memoryJev on the memory decisions in a coding agent.X post · cyrilXBT