New this weekPrompt Drop— visitors · 7d— submittedGet featured
0

Selvedge

In short: Selvedge is an open-source coding AI tool. Free, MIT-licensed command-line tool and local MCP server that gives AI coding agents a memory of past decisions. X Buzz Score: 3/100 (1 posts in 7 days).

Verdict: Free, MIT-licensed command-line tool and local MCP server that gives AI coding agents a memory of past decisions. You or your agent save why an approach was chosen, what was rejected and what would change your mind, tied to a function, database column, API route or dependency. In the next session the agent can look that up before editing. Everything stays in a SQLite file in your repo's .selvedge folder, with no account and no LLM calls. It works with any agent that supports MCP or has shell access. Install it with uv (Python 3.11+). There is also an optional GitHub Action that posts recorded reasons on pull requests. It is a small project (v0.3.18, about 26 GitHub stars) and only remembers what you save; it doesn't check whether those notes are still true. Submitted by the maintainer; passed our read-only safety check (public MIT repo since Apr 2026, PyPI releases, docs site).

Use this if you switch between Claude Code, Codex or Cursor sessions and keep re-explaining why the code is the way it is.

c/codingc/agentsopen-source

Visit Selvedge →

Live X Buzz Score

X 3/100 Quiet updated 17 Nov, 09:28 HKT

0posts, last 24h
1posts, last 7 days
0×24h vs prior daily avg (under 1)
0likes in top 0 sampled posts

Daily X posts, last 7 days (oldest → today): 0 · 0 · 0 · 1 · 0 · 0 · 0. Score parts: volume 3/50, spike 0/25, engagement 0/25. Method

No standout X posts yet (we only show posts with 100+ likes or 10K+ views).

Alternatives

  • claude-mem: Apache-2.0 plugin that gives coding agents memory across sessions: it records what the agent did, compresses it with AI, and injects the relevant bits into the next session. Works with Claude Code, Codex, Gemini, Copilot, OpenCode and more. v15.9.0 shipped Nov 2 2026; ~1.1K stars in a day on GitHub trending Nov 4 (~112K total). Stored session logs can include code and secrets, so check what it keeps.
  • Context Mode: Sits between coding agents and the model: folds tool output, keeps session memory, and blocks commands by rules you set.
  • GBrain: Garry Tan's MIT-licensed memory layer for the agent you already use: it stores facts with their sources, lets you correct or withdraw them, and serves the same memory to Codex, Claude Code, Hermes, OpenClaw and other MCP clients. Starts keyless with keyword search; semantic search and background enrichment are optional add-ons. GitHub garrytan/gbrain has about 32K stars and was on GitHub Trending on Nov 10, 2026; gbrain.io is optional paid hosting. Self-hosting is a developer setup (CLI, MCP, Tailscale for remote access), and its opinions are one person's workflow.

Similar AI tools

  • Gemini 4 Neon: Google's 18 Nov 2026 Gemini 4 model, aimed at fast end-to-end software engineering, knowledge work and science. API price is $3 in / $15 out per million tokens ($0.15 cached input) with a 1M-token context window. It is rolling out in the Gemini API and AI Studio on the paid tier, in Gemini for business, and to Google AI Pro and Ultra subscribers in Antigravity and the Gemini app — not described as a free-tier default. Google reports a 61 on the Artificial Analysis Intelligence Index at its xhigh thinking level and #1 on that firm's Coding Agent Index; treat those as Google's claims. In Antigravity it can also orchestrate parallel Gemini 4 Hydrogen subagents.
  • Claude: Excellent for long documents, writing and careful coding help.
  • OpenCode 2.0: Open-source, model-agnostic terminal coding agent with a native TUI and multi-session support.
  • GPT-6.2 Astra: OpenAI's new top model for hard professional work, released Nov 17, 2026. OpenAI says it makes slides, spreadsheets and documents that follow your templates more closely and 'read less like AI', and that it writes cleaner code, especially for websites, games and 3D design. It's in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users (not the free plan), and in the API as gpt-6.2-astra at $8 in / $40 out per million tokens, 20% less than GPT-6 Astra. Artificial Analysis puts it second on its Intelligence Index, just behind Claude Fable 5.2. Early testers on X report mixed results on long tasks, and OpenAI's system card says it was more likely than its predecessor to leave out obstacles when reporting its work, so check what it tells you it did.
  • GPT-6.2 Sol: OpenAI's faster, cheaper everyday model, released Nov 17, 2026 alongside GPT-6.2 Astra. OpenAI says it is smarter and more token-efficient than GPT-6.1 Sol, handles multistep agent workflows well, and that many OpenAI staff use it as their default. It's in ChatGPT Work and Codex for Plus, Pro, Enterprise and Edu users, and Pro 200 subscribers get the Ultrafast mode with Sol for the first time. In the API it's gpt-6.2-sol at the same price as 6.1 Sol: $2 in / $10 out per million tokens ($0.10 cached input), about a fifth of Astra's token price. The quality claims are OpenAI's own.
  • ChatGPT: The default general assistant; strong all-rounder with a huge plugin and GPT ecosystem.

See all: best AI tools for coding · trending AI tools today

FAQ

What is Selvedge?

Free, MIT-licensed command-line tool and local MCP server that gives AI coding agents a memory of past decisions. You or your agent save why an approach was chosen, what was rejected and what would change your mind, tied to a function, database column, API route or dependency. In the next session the agent can look that up before editing. Everything stays in a SQLite file in your repo's .selvedge folder, with no account and no LLM calls. It works with any agent that supports MCP or has shell access. Install it with uv (Python 3.11+). There is also an optional GitHub Action that posts recorded reasons on pull requests. It is a small project (v0.3.18, about 26 GitHub stars) and only remembers what you save; it doesn't check whether those notes are still true. Submitted by the maintainer; passed our read-only safety check (public MIT repo since Apr 2026, PyPI releases, docs site).

Who should use Selvedge?

Use Selvedge if you switch between Claude Code, Codex or Cursor sessions and keep re-explaining why the code is the way it is.

Is Selvedge free?

Yes. Selvedge is open source, so you can use and self-host it for free. Any hosted version or the AI model it calls may cost extra.

What are alternatives to Selvedge?

Alternatives include claude-mem, Context Mode, GBrain, Gemini 4 Neon.

Get your badge: maker of Selvedge? Add a free "Featured on toolhype" badge
Featured on toolhype badge (light)Featured on toolhype badge (dark)

HTML (light)

<a href="https://toolhype.dev/tools/selvedge" target="_blank"><img src="https://toolhype.dev/badge/featured-light.svg" alt="Selvedge is featured on toolhype" width="220" height="54" /></a>

Markdown (light)

[![Selvedge is featured on toolhype](https://toolhype.dev/badge/featured-light.svg)](https://toolhype.dev/tools/selvedge)

HTML (dark)

<a href="https://toolhype.dev/tools/selvedge" target="_blank"><img src="https://toolhype.dev/badge/featured-dark.svg" alt="Selvedge is featured on toolhype" width="220" height="54" /></a>

Markdown (dark)

[![Selvedge is featured on toolhype](https://toolhype.dev/badge/featured-dark.svg)](https://toolhype.dev/tools/selvedge)

About the badge

✉️ Never miss the next hot toolWeekly: what's hyped on X, and what's worth it.

Discussion

Comments are coming with sign-in.

Amanbehumble avatar
Amanbehumble@amanbehumble

Daily AI tools & vibe-coding drops.

21K followers · as of Oct 16

✉️ The weekly AI tools dropSubscribe