Learn to build a TypeScript CLI research agent with specialized sub-agents for deep research, auditing, and writing. Features Docker, Modal GPU, and Node.js.

It’s an architectural solution to a prompting problem. By giving each agent a smaller, more focused task, the quality of the output skyrockets because the LLM isn't trying to juggle a dozen different goals at once.
Build a TypeScript CLI research agent with four specialized sub-agents: Researcher, Reviewer, Writer, and Verifier. Deliver capabilities as Markdown skill files synced to a local skills directory on startup. Expose slash commands (/deepresearch, /lit, /audit, /replicate) that dispatch agents in parallel across papers, web, and repos. Ground every output in cited sources with direct URLs. Add Docker for safe execution and Modal/RunPod for GPU compute. Stack: TypeScript, Node.js.


The system utilizes four distinct sub-agents: Researcher, Reviewer, Writer, and Verifier. These agents work in parallel to handle complex tasks such as deep research, literature reviews, and code replication. By delivering capabilities as Markdown skill files synced to a local directory, the CLI ensures that each agent has a specific, grounded role in the research workflow, providing comprehensive outputs backed by cited sources and direct URLs.
The research agent uses specialized slash commands like /deepresearch and /audit to dispatch agents across web sources, academic papers, and repositories. The Researcher agent gathers data while the Verifier ensures all claims are grounded in cited sources. This multi-agent architecture, built on Node.js and TypeScript, allows for parallel processing of information, ensuring that every output is both factually accurate and thoroughly vetted before being presented to the user.
To ensure a secure and high-performance environment, the project integrates Docker for safe code execution and sandboxing. For intensive computational tasks, it leverages Modal and RunPod for scalable GPU compute. This stack, combined with a TypeScript and Node.js foundation, allows the CLI research agent to handle demanding workloads like repository analysis and complex simulations without compromising the security of the local host environment.
Yes, the system is designed to be highly extensible through Markdown skill files. Upon startup, the CLI syncs these files to a local skills directory, allowing users to define new behaviors or refine existing agent roles. By using the provided slash commands, you can trigger these specialized skills to perform targeted actions across different data sources, making it a flexible tool for developers and researchers looking to automate their deep research workflows.
Cree par des anciens de Columbia University a San Francisco
"Instead of endless scrolling, I just hit play on BeFreed. It saves me so much time."
"I never knew where to start with nonfiction—BeFreed’s book lists turned into podcasts gave me a clear path."
"Perfect balance between learning and entertainment. Finished ‘Thinking, Fast and Slow’ on my commute this week."
"Crazy how much I learned while walking the dog. BeFreed = small habits → big gains."
"Reading used to feel like a chore. Now it’s just part of my lifestyle."
"Feels effortless compared to reading. I’ve finished 6 books this month already."
"BeFreed turned my guilty doomscrolling into something that feels productive and inspiring."
"BeFreed turned my commute into learning time. 20-min podcasts are perfect for finishing books I never had time for."
"BeFreed replaced my podcast queue. Imagine Spotify for books — that’s it. 🙌"
"It is great for me to learn something from the book without reading it."
"The themed book list podcasts help me connect ideas across authors—like a guided audio journey."
"Makes me feel smarter every time before going to work"
Cree par des anciens de Columbia University a San Francisco
