Fetching from the wire…
Public story · 2026-03-04 · source-backed
| Project | Stars | What It Does | Why It Matters |
|---|---|---|---|
| Worktrunk | 2.7K | Rust CLI for Git worktrees with AI agents | Auto-squash/rebase/merge, LLM commit messages, build cache sharing. From PRQL creator. |
| Timber | 545 | Compile XGBoost/sklearn to C99 binaries | 336x faster than Python, 48KB artifacts, MISRA-C compliant. Ollama for classical ML. |
| Codebuff | 3.6K | Open-source multi-agent coding tool | Claims 61% vs Claude Code's 53% on 175-task eval. Any model via OpenRouter. |
| Hive Memory | new | MCP cross-project memory for agents | Fully local, cross-project context sharing. Show HN featured. |
| Sub-500ms Voice Agent | 562 HN | Production voice agent in one day for $100 | Groq ~80ms TTFT + streaming pipeline = 2x faster than Vapi. Blueprint documented. |
| PDF Oxide | 304 | 5x faster PDF processing than PyMuPDF | Rust core, Python/JS/WASM bindings. MCP server included. |
| Omni | 503 | Self-hosted workplace search (open-source Glean) | Unified search across Google Workspace, Slack, Jira. ParadeDB single-Postgres. |
Each link below shares sources, entities, or timing with this story.
Stripe published Part 2 of its Minions engineering blog, and it's the most detailed production agent architecture I've read from any company this year. The numbers alone are worth the read: 1,300+ weekly merged PRs from coding agents. But the architecture decisions matter more...
AOT compiler turning XGBoost, LightGBM, scikit-learn models into native C99 inference code. 336x faster than Python inference. Created Feb 27 — very fresh. GitHub
Three moves, two days, no coordination between them. August 10–11: GitHub shipped Ollama as a BYOK provider inside Copilot for JetBrains (GitHub Changelog). Unsloth released Unsloth Desktop with a command literally named unsloth start claude, which points Claude Code and Codex...
akitaonrails/ai-memory (Rust, 2,485 stars) hooks into agent lifecycles to capture prompts, tool calls, and session boundaries, compiling them into a git-versioned per-project markdown wiki with full-text search and optional vector retrieval. Pitch: abandon a task in Claude Cod...
Girder parses a repo into functions, definitions and call edges, then serves exact function source, callers and callees, impact analysis, minimal test selection and graph-addressed edits over MCP as one static Rust binary. Its committed ten-node measurement returns 8,765 bytes...
On August 25, Ninjō AI shipped AI sales agents for Instagram and WhatsApp built and managed through MCP from Claude Code, ChatGPT and Codex; coolplugz shipped a Claude Code orchestrator pulling context from Jira, GitHub, Notion and Slack; Diet Claude shipped a usage meter for...
MindPattern daily
One email a day at 7 AM. Sources and a take on every story. Unsubscribe anytime.