OSS Tanbou
View on GitHubProject website

Letta Code — stateful agents that learn through memory, identity, and skills

OSS scale score 208.4OSS health 89Update pace 342 commits/30d
About these scores

OSS scale score is an unbounded metric that log-compresses and weights Stars, Watchers, Forks, and Contributors. Discovery score is the current OSS scale score minus the score at discovery. Update pace is commits in the last 30 days, growth momentum is the OSS scale score difference within the recent observation window, and OSS health is a 0–100 rating based on available recency, Community Health, and release data.

Stars
3,385
Primary language
TypeScript
License
Apache-2.0
Repository last updated
Sep 20, 2026

Overview

Letta Code is a stateful agent harness for agents with persistent memory, identity, and skills. The same agent can be accessed from the CLI, desktop, browser, Slack, Telegram, Discord, and custom channels while evolving over time.

Features and best fit

Based on official documentation; not hands-on tested · Content checked:

Let agents update their own memory and skills

Memory blocks and skills can be rewritten over time so prior experience becomes part of persistent context.

Sources: [2][3]

Combine MemFS, subagents, channels, hooks, permissions, and schedules

Git-tracked memory, subagents, messaging channels, cron schedules, and execution controls are exposed through one agent harness.

Sources: [2][3]

For long-lived agents that need continuity beyond a single task

It fits agents that should retain identity and experience while operating across multiple machines and communication channels.

Sources: [2]

Separate cloud-backed capabilities from local ones

Some capabilities, including remote computers and secrets, require signing in to Letta. Decide where agent state and credentials should live.

Sources: [2][3]

Official sources

  1. [1]letta-ai/letta-code repository(2026-09-20)
  2. [2]Letta Code README(2026-09-20)
  3. [3]Letta Code documentation(2026-09-20)
  4. [4]Letta Code Apache-2.0 license(2026-09-20)
Supplemental curator note

Letta Code is designed around persistent agents that rewrite memory and skills over long time horizons rather than one-shot coding sessions. Some features such as remote computers and secrets require signing in to Letta, so distinguish local capabilities from cloud-backed ones.

Try it in 3 steps

  1. 1

    Get the source

    git clone --depth 1 https://github.com/letta-ai/letta-code.git
  2. 2

    Enter the repository

    cd letta-code
  3. 3

    Check the official steps

    Continue with the commands in the README Installation, Quick Start, or Getting Started section.

    find . -maxdepth 1 -iname 'README*' -exec sed -n '1,220p' {} \; -quit
Check the official README

Growth

Growth trends · Last 30 days

3,385 Stars

Trend data is still being collected.

Development activity

Last 90 days · weekly

Commits (last 30 days)
342
Open PRs
238

Development activity is still being collected.

Built with

Categories and tags

Categories

Tags

GitHub data

GitHub dataView detailed GitHub data
View on GitHub

GitHub Topics

  • agent-memory
  • ai
  • claude
  • codex
  • continual-learning
  • letta
  • memgpt
  • stateful-agents
Stars
3,385
Forks
414
Watchers
13
Open issues
142
Contributors
51
Primary language
TypeScript
License
Apache-2.0
Repository last updated
Sep 20, 2026
Write a related article

Share a guide or use case for this OSS in Markdown. Articles are published after administrator approval.

Report incorrect information

Tell us if any listing information is incorrect or outdated.