Nodecode is an agent that rewrites itself. It runs as one process in SBCL, a Common Lisp system, and shells attach to it. The model has a single tool: it evaluates Lisp in the running program and keeps what it defines. Told it cannot read PDFs and should teach itself, it defines read-pdf, keeps it, and reads the invoice. First launch opens a setup that names any other coding agent already on the machine — Claude Code, Codex, Hermes Agent, OpenCode, Cursor CLI, Gemini CLI and more — and asks before reading anything; say Import and it brings over that agent's keys, instructions, memory, skills, MCP servers, bots and jobs.
Be the first to rate Nodecode. Your feedback helps other users decide.
Tell it what it cannot do; it defines the function, the running program keeps it, and it uses it next time.
Let setup find your current coding agent and carry over keys, instructions, memory, skills and MCP servers.
Connect ordinary shells to the one long-running process and keep working in the same image.
Similar tools in the same category — compare side-by-side
ADRs4AI keeps the reasoning behind human and AI work instead of letting it evaporate with the sessio...
Zephium is an open-source browser and work environment built for humans and agents on the same surfa...
Jukebox OS is a deterministic flight recorder, ACID tool sandbox and time-travel runtime for multi-a...
Guides and comparisons about code tools
Nine launches where the agent is pointed at law, hiring, money, packaging, language learning and images instead of a repository.
Ten recent launches for running agents where they can do less harm: virtual machines, air-gapped sandboxes, sequencers, and control planes you host yourself.
Ten launches aimed at the same bottleneck: agents re-read their own history, and every re-read is billed. Compression, code mode, prices and measurement.
Ten recent MCP servers that connect agents to data and systems they cannot invent: B2B records, financial metrics, flights, CAD, and keyword volumes.
Five Show HN launches about where agents run: an incident benchmark, a policy guardrail, a container orchestrator, a one-megabyte C agent and a 3D code world.
