ai-agent

recursive-llm

Name: recursive-llm
Author: grishahq

Recursive Language Models for efficient long-context processing. Analyze 1M+ tokens by storing context in a Python REPL while reducing LLM token usage.

614 stars+2★ this week87 forksPythonUpdated 10/2/2026100% free · open source

What it does

The recursive-llm tool enables efficient long-context processing by storing context in a Python REPL, allowing analysis of over 1 million tokens while reducing language model token usage.

When to use it

•When you need to process large texts or conversations that exceed the token limits of traditional language models
•When you want to reduce the cost of using language models by minimizing token usage
•When you need to maintain context across multiple interactions or turns in a conversation

Ready-to-paste prompt

rlm.process('Analyze the following text: ' + ' '.join([str(i) for i in range(100000)]))

Heads up: Make sure you have a compatible Python version (3.8 or later) and sufficient memory to handle the recursive context storage

Saves to your device

Use with Claude

New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

🔑 Needs an API key

Works after you add credentials — the setup agent will ask you for them.

Try it instantly — no install

Claude Code

mkdir -p ~/.claude/skills/recursive-llm && curl -fsSL https://workflowstacks.com/api/skills/recursive-llm/claude-skill -o ~/.claude/skills/recursive-llm/SKILL.md

Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)

claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

How recursive-llm works

Codeflow

Free to inspect

Recursive-llm is a large Python project (~15k lines across 102 code files, plus 23 test files). Setup is light: installs like a normal app — add your API keys. Reading the code is optional. Last commit a month ago, MIT license, has a test suite.

Size

Large codebase

~15k lines · 102 code files · ~2 h to skim

Setup

One-command install

Installs like a normal app — add your API keys. Reading the code is optional.

Runs on

Python

Needs API keys (.env)

Python 48%TypeScript 44%CSS 7%JavaScript 1%

Where to start reading

1
README.md
Start here — what it does and how to install it
2
pyproject.toml
Dependencies and the commands it exposes
3
.env.example
The API keys and settings you must provide
4
src/rlm/__init__.py
Inside src/ — the main logic begins here
5
examples/basic_usage.py
A worked example — copy this to get going

What's in each folder

src/Core code — the actual logic12 files

examples/Examples you can copy12 files

visualizer/Folder93 files

experiments/Research & notebooks27 files

benchmarks/Evaluations & benchmarks7 files

tests/Tests — proof it works21 files

.github/CI / automation (GitHub Actions)3 files

READMEHas testsDocumentedCI checksExamples includedMIT licenseUpdated this month