{"name":"RLM: Process Large Documents in LLM-Sized Chunks","description":"RLM is a .NET CLI for breaking large documents into manageable chunks, processing them across multiple turns, and combining the results. It suits developers using Claude Code or scripts to work with content that exceeds an LLM's context window.","github":"https://github.com/endjin/RLM","url":"https://osrepos.com/repo/endjin-rlm","source":"osrepos.com","sourceDescription":"This repository profile is provided by osrepos.com, an open source repository discovery platform.","repositoryProfile":"https://osrepos.com/repo/endjin-rlm","generatedFor":"open source discovery and AI-assisted research","markdown":"https://osrepos.com/repo/endjin-rlm.md","json":"https://osrepos.com/repo/endjin-rlm.json","topics":["csharp","cli","ai-agents","claude-code","document-chunking","recursive-language-models"],"keywords":["csharp","cli","ai-agents","claude-code","document-chunking","recursive-language-models"],"stars":null,"summary":"RLM is a .NET CLI for breaking large documents into manageable chunks, processing them across multiple turns, and combining the results. It suits developers using Claude Code or scripts to work with content that exceeds an LLM's context window.","content":"## Overview\n\nRLM is a .NET command-line tool that helps agents and developers work with documents too large to fit in an LLM context window. It loads content from common document formats, decomposes it for iterative processing, and lets callers store and combine partial results.\n\nThe tool is aimed at workflows where context management and repeatable chunk processing matter, rather than at providing an LLM or an end-to-end document analysis service. Its Claude Code skill and worker agent offer a ready-made integration path, while session and piping features support custom scripts and recursive workflows.\n\n## Key Features\n\n- Loads Markdown, PDF, HTML, JSON, Word documents, plain text, directories, and standard input.\n- Offers uniform, filtering, semantic, token-based, recursive, and query-based automatic chunking.\n- Maintains persistent session state for multi-turn processing.\n- Provides commands to inspect, navigate, filter, skip, and retrieve document sections or chunks.\n- Stores partial results and imports results from child sessions or external files for aggregation.\n- Includes a Claude Code skill and parallel chunk-processing worker.\n- Supports JSON and raw output for scripting and piping.\n\n## Use Cases\n\n- A Claude Code user needs to summarize a long report in stages and combine the chunk-level summaries.\n- A developer wants to search large document collections for matching text without loading everything into one prompt.\n- An agent workflow needs to split a complex document into child tasks, process them in isolated sessions, and import the results.\n- A team building a document-processing script needs consistent chunking and token-aware sizing through a CLI.\n\n## Our Take\n\nRLM is a focused fit for .NET and Claude Code workflows, but its quiet recent activity and highly concentrated contributor base merit a maintenance check before it becomes a critical dependency.\n\n**Good fit if:**\n- You need to process oversized documents through staged LLM or agent workflows.\n- Your tooling can use a .NET 10 CLI, and you value its documented Claude Code integration or script-friendly output.\n- You can evaluate and maintain a small project whose development is concentrated mostly in one contributor, while benefiting from its recent releases.\n\n**Look elsewhere if:**\n- Your environment cannot run .NET 10 or you need a service/API rather than a command-line workflow.\n- You require evidence of active commits and issue or pull-request resolution: the project is marked quiet, with no commits or closed issues in the measured recent period.\n- Your adoption policy requires a broadly distributed maintainer base; most contributions come from one person.\n\n## Project Health\n\n| Signal | Value |\n|---|---|\n| Status | **Quiet**: no commits in the last 90 days |\n| Last commit | 2026-04-21 (6 months ago) |\n| Commits, last 90 days | 0 |\n| Releases, last 12 months | 7 (latest 1.0.6, 2026-01-31) |\n| Contributors | 2 (top contributor: 94% of commits) |\n| Issues closed, last 90 days | 0 |\n| Pull requests merged, last 90 days | 0 |\n| Project age | 9 months |\n\nChecked on 2026-10-13 with the GitHub API.\n\n## Project Facts\n\n- Language: C#\n- License: Apache-2.0\n- Stars: 6\n- Forks: 1\n- Topics: ai, ai-agents, claude-code, context-engineering, rlm\n- Archived: no\n\n## What You Need\n\nDetected in the repository:\n\n- C# (main language on GitHub)\n- Automated checks on GitHub Actions\n\n## Getting Started\n\nInstall the global tool with .NET 10 or later:\n\n```bash\ndotnet tool install -g rlm\n```\n\nFor Claude Code setup, build instructions, and workflow examples, see the [README](https://github.com/endjin/RLM#readme).\n\n## License in Plain Words\n\n**Apache-2.0** (permissive).\n\n- Commercial use: yes\n- Modify and redistribute: yes\n- You must keep: the license, the NOTICE file and a note of your changes\n- Share your changes: no\n- Includes an explicit patent grant from the contributors\n\nA summary, not legal advice: the LICENSE file is what applies.\n\n## Considerations\n\n- The project is marked quiet, with no commits in the last 90 days. It has published releases in the last year, but measured issue and pull-request handling was also inactive in that period.\n- There are two contributors, and the contribution share is highly concentrated in one person. Teams relying on the tool should assess whether they can maintain or fork it if needed.\n- The CLI targets .NET 10 or later, which may be a constraint for teams that have not adopted that runtime.\n- Chunking and result aggregation organize work for an LLM workflow; the README does not describe an included model or hosted processing service. You need to supply the processing step separately.\n- The README documents a test project, but the repository health data does not report tests as a project requirement signal. Verify the test and CI coverage that matters to your own adoption before relying on it.","metrics":{"detailViews":1,"githubClicks":1},"dates":{"published":null,"modified":"2026-10-13T19:38:23.000Z"}}