[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-73851":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":23,"hasPages":23,"topics":25,"createdAt":10,"pushedAt":10,"updatedAt":26,"readmeContent":27,"aiSummary":28,"trendingCount":16,"starSnapshotCount":16,"syncStatus":29,"lastSyncTime":30,"discoverSource":31},73851,"letta-code","letta-ai\u002Fletta-code","letta-ai","The memory-first coding agent","https:\u002F\u002Fdocs.letta.com\u002Fletta-code",null,"TypeScript",2707,301,8,130,0,26,85,253,78,29.44,"Apache License 2.0",false,"main",[],"2026-06-12 02:03:18","# Letta Code\n\n[![npm](https:\u002F\u002Fimg.shields.io\u002Fnpm\u002Fv\u002F@letta-ai\u002Fletta-code.svg?style=flat-square)](https:\u002F\u002Fwww.npmjs.com\u002Fpackage\u002F@letta-ai\u002Fletta-code) [![Discord](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdiscord-join-blue?style=flat-square&logo=discord)](https:\u002F\u002Fdiscord.gg\u002Fletta)\n\nLetta Code is a memory-first coding harness, designed for long-lived agents that can learn from experience.\n\nInstead of working in independent sessions, you work with a persisted agent whose memory is portable across models (Claude, GPT, Gemini, GLM, Kimi, and more).\n\nRun Letta Code in the [**CLI**](https:\u002F\u002Fdocs.letta.com\u002Fletta-code\u002Fcli), or download the [**desktop app**](https:\u002F\u002Fdocs.letta.com\u002Fletta-code\u002Fdesktop-app) for MacOS, Windows, and Linux.\nYou can also access Letta Code via [your phone](https:\u002F\u002Fdocs.letta.com\u002Fletta-code\u002Fremote-mobile) and [Slack\u002FTelegram\u002FDiscord](https:\u002F\u002Fdocs.letta.com\u002Fletta-code\u002Fchannels).\n\n![](https:\u002F\u002Fgithub.com\u002Fletta-ai\u002Fletta-code\u002Fblob\u002Fmain\u002Fassets\u002Fletta-code-demo.gif)\n\n## Get started\nInstall the package via [npm](https:\u002F\u002Fdocs.npmjs.com\u002Fdownloading-and-installing-node-js-and-npm):\n```bash\nnpm install -g @letta-ai\u002Fletta-code\n```\nNavigate to your project directory and run `letta` (see various command-line options [on the docs](https:\u002F\u002Fdocs.letta.com\u002Fletta-code\u002Fcommands)). \n\nRun `\u002Fconnect` to configure your own LLM API keys (OpenAI \u002F ChatGPT, Anthropic, zAI coding plan, etc.), and use `\u002Fmodel` to swap models.\n\nFor slow local inference servers, configure a provider-level timeout when connecting. For example, LM Studio-compatible llama-server backends that need up to 10 minutes for large-context compaction can use:\n\n```bash\nletta --backend local connect lmstudio --base-url http:\u002F\u002F127.0.0.1:1234\u002Fv1 --timeout 600s\n```\n\nTimeouts are stored per local provider in milliseconds; pass `--no-timeout` or `--timeout false` to disable the provider timeout.\n\nYou can also download the [**desktop app**](https:\u002F\u002Fdocs.letta.com\u002Fletta-code\u002Fdesktop-app) for MacOS, Windows, and Linux. Agents created in the CLI are available via the desktop app, and vice versa.\n\n## Philosophy \nLetta Code is built around long-lived agents that persist across sessions and improve with use. Rather than working in independent sessions, each session is tied to a persisted agent that learns.\n\n**Claude Code \u002F Codex \u002F Gemini CLI** (Session-Based)\n- Sessions are independent\n- No learning between sessions\n- Context = messages in the current session + `AGENTS.md`\n- Relationship: Every conversation is like meeting a new contractor\n\n**Letta Code** (Agent-Based)\n- Same agent across sessions\n- Persistent memory and learning over time\n- `\u002Fclear` starts a new conversation (aka \"thread\" or \"session\"), but memory persists\n- Relationship: Like having a coworker or mentee that learns and remembers\n\n## Agent Memory & Learning\nIf you’re using Letta Code for the first time, you will likely want to run the `\u002Finit` command to initialize the agent’s memory system:\n```bash\n> \u002Finit\n```\n\nOver time, the agent will update its memory as it learns. To actively guide your agents memory, you can use the `\u002Fremember` command:\n```bash\n> \u002Fremember [optional instructions on what to remember]\n```\nLetta Code works with skills (reusable modules that teach your agent new capabilities in a `.skills` directory), but additionally supports [skill learning](https:\u002F\u002Fwww.letta.com\u002Fblog\u002Fskill-learning). You can ask your agent to learn a skill from its current trajectory with the command: \n```bash\n> \u002Fskill [optional instructions on what skill to learn]\n```\n\nRead the docs to learn more about [skills and skill learning](https:\u002F\u002Fdocs.letta.com\u002Fletta-code\u002Fskills).\n\nCommunity maintained packages are available for Arch Linux users on the [AUR](https:\u002F\u002Faur.archlinux.org\u002Fpackages\u002Fletta-code):\n```bash\nyay -S letta-code # release\nyay -S letta-code-git # nightly\n```\n\n---\n\nMade with 💜 in San Francisco\n","Letta Code 是一个以记忆为中心的编码代理，专为能够从经验中学习的长期存在的代理设计。其核心功能包括持久化的代理记忆，该记忆可以在不同模型间移植（如Claude、GPT、Gemini等），并且支持通过CLI、桌面应用程序以及手机和多种即时通讯平台访问。技术特点上，Letta Code使用TypeScript编写，确保了跨平台的一致性和性能。它特别适合需要持续迭代开发环境下的软件工程师使用，尤其是在希望构建能够随时间累积知识并提高效率的编码助手时。此外，用户可以通过配置自定义的LLM API密钥来连接不同的后端服务，增加了使用的灵活性。",2,"2026-06-11 03:47:39","high_star"]