frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

fp.

Open in hackernews

Show HN: Darc – grep-like memory search tool for coding agents

https://github.com/0xjunha/darc
1•jnhprk•1h ago
Hi HN, I’m Junha Park. I've been experimenting with agent memory, especially how to make agents run more reliably on large tasks.

I built Darc, an open-source shared memory search tool for coding agents, with a different approach from most agent memory systems we can see today. It is an index + (lexical) search tool over agent session history, rather than a managed memory system. No embeddings, no agent-aided consolidation, no injection hooks.

How it works: Darc archives Codex / Claude Code session rollouts that already exist under ~/.codex and ~/.claude, indexes them into a single SQLite DB, and exposes session/turn/tool call/file-level search commands over them.

Why I built this: I've tried using different agent memory tools and they were useful, but sometimes I found them nosiy. I kept turning memory on/off depending on the task. For example, when I ran iterative code review rounds, I saw reviewer agents report "no findings" with "memory cited" memos, where the cited memory came from recent review sessions or the session that built the feature being reviewed. I believe this could put bias to the current reviewer's context window, and wanted to turn off those "injection hook" memory feature for such cases.

So I came up with an idea. Recent coding agent tools tend do prefer simple lexical search (UNIX tools; `rg`, `sed`, etc.) to semantic search over vector embeddings, and they work quite well and reliably without the heavy embedding. Why don't we take similar approach to agent memories, treating the session history like documents or code?

The idea is to let agents use Darc iteratively to discover relevant evidence or decisions from prior sessions, see which files or sessions are deeply related, and keep chaining the search until they understand the full picture of what happened in the past before they jump into editing code.

I also added team sharing feature to Darc. Users can encrypt (via `age`) and share their indexed agent session history with other team members working on the same project via Git backend of their choice (GitHub or self-hosted Git server). Regex-based redaction is supported by default (so sensitive data like secret keys, API keys or env vars don't get into SQLite at all). Then, users can pull agent session history indexes from other team members and search over them to get decision context made by others. This is analogous to asking colleagues who wrote the code about context behind the work, like design decisions, before touching that part of the code.

That said, it still needs proper evaluation. I'm working on simple bench/evals to compare scenarios like:

- Baseline, with no memory

- Using Codex / Claude Code built-in memory (the native memory system that summarizes prior sessions using background agents and writes MEMORY.md, not merely AGENTS.md or CLAUDE.md)

- Using Darc

- Using Darc + Codex / Claude Code built-in memory

- Just `rg` on session history directly

If this approach sounds interesting (or you think it doesn't make sense) I'd love to hear your feedback. I'd also love to learn what methods people are using to manage team-level context on project where coding agents are used heavily.

Waymo pauses Atlanta service as its robotaxis keep driving into floods

https://techcrunch.com/2026/05/21/waymo-pauses-atlanta-service-as-its-robotaxis-keep-driving-into...
1•mattas•2m ago•0 comments

Ask HN: Danger or Fun? encoding secret messages into HN comments

1•smalltorch•2m ago•0 comments

The Amish Are Falling in Love with AI

https://nymag.com/intelligencer/article/why-the-amish-are-falling-in-love-with-ai.html
1•bookofjoe•3m ago•1 comments

The World Needs Your Great Work

https://www.palladiummag.com/2026/03/17/the-world-needs-your-great-work/
1•ronfriedhaber•3m ago•0 comments

Wes mckinney releases multiple bangers

https://kenn.io/
1•pbd•3m ago•0 comments

Neanderthals may have been doing dentistry 59,000 years ago – without anesthetic

https://www.cbc.ca/radio/asithappens/neanderthals-may-have-been-dabbling-in-dentistry-59-000-year...
1•Brajeshwar•5m ago•0 comments

Brazil bans stablecoin and crypto settlement in cross-border payments

https://www.coindesk.com/policy/2026/05/02/brazil-s-central-bank-bans-stablecoin-and-crypto-settl...
1•sevenfoldnancy•5m ago•0 comments

Spotify Will Start Reserving Concert Tickets for Fans

https://www.hollywoodreporter.com/music/music-industry-news/spotify-will-start-reserving-concert-...
1•elffjs•6m ago•0 comments

Show HN: Write your BPF programs in Go, not C

https://github.com/boratanrikulu/gobee
1•boratanrikulu•7m ago•0 comments

Vitamin D acts more like a hormone than a vitamin

https://www.empirical.health/blog/vitamin-d-is-a-hormone/
1•brandonb•8m ago•0 comments

How Semantic SQL Works

https://cube.dev/blog/how-semantic-sql-works
1•pavel_tiunov•8m ago•0 comments

We replaced Redis with MySQL for inventory reservations – and it scaled

https://shopify.engineering/scaling-inventory-reservations
2•birdculture•8m ago•0 comments

1986: Email – The perfect tech for the Jet Set? BBC [video]

https://www.youtube.com/watch?v=JqJ159pngY8
2•ljoshua•9m ago•0 comments

Project Hail Mary – Stellar Navigation Chart

https://valhovey.github.io/gaia-mary/
3•speleo•10m ago•0 comments

Mnemo – mobile reading companion that reflects with you

https://trymnemo.net
1•owenHochwald•10m ago•0 comments

Gemini 3.5 deleted 28,745 lines, broke production, and wrote a fake post-mortem

https://old.reddit.com/r/Bard/comments/1tisrg1/gemini_35_deleted_28745_lines_broke_production/
1•YeGoblynQueenne•10m ago•0 comments

Zero Lang

https://github.com/vercel-labs/zerolang
2•https443•11m ago•0 comments

Fixing a 20 year old bug in Debian curl

https://blog.sergiodj.net/posts/fixing-20-year-old-bug-debian-curl/
1•speckx•12m ago•0 comments

Inside the AI compute crunch driving Google researchers to quit

https://www.latimes.com/business/story/2026-05-18/inside-ai-compute-crunch-driving-google-researc...
2•1vuio0pswjnm7•16m ago•0 comments

WebRTC Latency: side-by-side comparison [video]

https://www.youtube.com/watch?v=OmLS9rE6dTU
2•chfritz•16m ago•0 comments

Kansas City Public Schools goes all-in on Apple with 30k devices

https://macdailynews.com/2026/05/21/kansas-city-public-schools-goes-all-in-on-apple-transitioning...
3•speckx•17m ago•1 comments

A star gone rogue tears through the Galaxy

https://www.nature.com/articles/d41586-026-01591-9
4•Brajeshwar•19m ago•0 comments

Satcove – A consensus engine for important questions

https://satcove.com
1•abyssalg•19m ago•1 comments

Heretic has been served a legal notice by Meta, Inc

https://old.reddit.com/r/LocalLLaMA/comments/1tjmvx6/heretic_has_been_served_a_legal_notice_by_me...
3•miav•19m ago•0 comments

A critical initialization for biological neural networks

https://www.nature.com/articles/s41586-026-10528-1
2•bookofjoe•19m ago•0 comments

The Agentic Infrastructure Era

https://www.pulumi.com/blog/the-agentic-infrastructure-era/
1•adamgordonbell•20m ago•0 comments

Sonar – Web Audio search for AI Agents (podcasts, news, earnings calls) via API

https://sonarapi.com/
1•surendersingh•20m ago•0 comments

Windows Registry Adventure: Hives and the registry layout

https://projectzero.google/2024/10/the-windows-registry-adventure-4-hives.html
1•ankitg12•20m ago•1 comments

Show HN: Benchd: client-side WASM benchmarking in the browser

https://github.com/Userfrom1995/benchd
1•userfrom1995•21m ago•0 comments

Malaysia orders TikTok explain 'grossly offensive' fake content targeting king

https://apnews.com/article/malaysia-tiktok-king-offensive-content-a9327060bdd4e4ce90c5e6021b8637c7
2•1vuio0pswjnm7•21m ago•0 comments