Fetching from the wire…
OSS2026-07-25 · source-backed
1jehuang/jcode reached 11,328 stars (+2,773 this week) as a Rust terminal agent harness at v0.9.1888-dev with 6,010 commits, MIT. The README publishes head-to-head numbers: 27.8MB RAM at one session and 260.8MB at ten, versus 833-3,237MB for competing harnesses, and 14.0ms time-to-first-frame versus 383-3,437ms. It embeds every conversation turn as a semantic vector so recall happens via cosine similarity instead of token-heavy memory tool calls, plus swarm mode (multiple agents in one repo with conflict resolution) and a self-modification mode where the agent edits and reloads its own source. Publishing your own memory footprint against named competitors is a confidence move.
Each link below shares sources, entities, or timing with this story.
MemPalace (~53.6K stars) claims best-benchmarked, against mem0 (~57.8K) and claude-mem (~80.8K) (GitHub). Agent memory went from experimental nicety to a competitive subcategory with published benchmarks. Pair this with the local-first angle: Mnemo offers a Rust + SQLite + pet...
Most coding agents run on JavaScript runtimes and eat 300MB of RAM just sitting idle. Zerostack is built in pure Rust, uses ~8MB on idle, ~12MB while working, and posts 0.0% idle CPU. It hit 1.0 on crates.io this week with 474 points and 252 comments on Hacker News. The archit...
macro-inc/macro hit #4 on GitHub trending at 2,821 stars total, shipping email, chat, docs, tasks, calls and CRM that all @-link into a shared context graph agents read as team-level memory. Explicitly "fully open source, not open core," with commercial licensing sold separate...
~248 stars today, 1,251 total, Rust, AGPL-3.0, unifying email, chat, docs, tasks, agents, calls and CRM with @-linking across all of them backed by shared memory (GitHub). The claim is that agent usefulness is bounded by context fragmentation, so you collapse the tools instead...
This Rust harness (+2,585 stars) competes on resource footprint rather than features: 27.8 MB PSS for a single session with local embedding disabled, claimed 13.9× less than Claude Code and 6× less than jcode's own embedding-enabled mode. Time-to-first-frame 14.0ms against a c...
tt-a1i/archify is an MIT-licensed agent skill turning a repo or description into architecture, workflow, sequence, data-flow and lifecycle diagrams, exported as self-contained HTML plus PNG, SVG, WebM and a share card. Its verifiability claim is mechanical rather than rhetoric...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.