Agent-first cognitive substrate managing cross-session memory as a self-validating filesystem graph with 18 manifest-driven verbs.
Copy the AI prompt to install this server into Claude Code, Cursor, or another agent β or use 1-click editor setup below.
One-click editor setup isnβt available for this listing yet β we donβt have a confirmed install command, and weβd rather show nothing than point your editor at the wrong package or host. Follow the projectβs own setup instructions, linked above.
Inspect callable tools, capabilities, and parameters exposed to AI agents by Myco.
β οΈ Deprecated: Myco has been reborn as v0.9-genesis
This PyPI package (
myco, v0.4βv0.8.x) is the Agent-First cognitive-substrate embryo: a 20-verb tool framework for an AI agent's memory. It is doctrinally a dead embryo, sealed at the git tagv0.8.8-final-embryo.The project has been reborn as a different conception: a decades-long symbiotic partnership between one human and a substrate that carries memory, character, and doctrine across LLM model generations. v0.9 is not a
pip install. It is a clone-and-build substrate (Rust + Python + TypeScript).π See github.com/Battam1111/Myco
Branch
v0.9-genesis, with itsGETTING_STARTED.md. ThismycoPyPI package is no longer the project; it remains only as the predecessor lineage.
Devour everything. Evolve forever. You just talk.
ClaudeΒ CodeΒ /plugin install myco@myco Β Β·Β ClaudeΒ DesktopΒ myco-install host cowork Β Β·Β AnyΒ MCPΒ hostΒ pip install 'myco[mcp]'
What it is Β· How it lives Β· Quick start Β· Verbs Β· Self-validation
Languages: English Β· δΈζ Β· ζ₯ζ¬θͺ
LangChain. LangGraph. CrewAI. DSPy. Claude Code skills. OpenHands. OpenClaw. Every few months the next framework drops and you migrate again.
Your notes rot too. The API you read three weeks back has changed. The doc you wrote last year is wrong now. Your AI does not even remember last week's decisions. Every new conversation starts from zero.
Now imagine one living substrate. It ingests the frameworks, the papers, the APIs, the codebases, the datasets, the decisions, the frictions. It keeps them connected in a graph the agent actually reads. It catches its own drift. It reshapes itself when your work outgrows its old form. For six months. For six years. No migration.
Myco is a living cognitive substrate for your AI agent.
Everything the agent reads or writes, every piece of code and every paper and every decision and every friction, lives on your filesystem as plain markdown + YAML. Files cross-reference each other by name, forming a graph the agent actually reads. The agent eats raw material, digests it into integrated knowledge, immune-checks its own work against drift, propagates learning across your projects, and when the shape of the substrate no longer fits the shape of the work, reshapes the substrate itself. The kernel that runs all of this is itself a substrate, editable by default, maintained by the same agent that uses it.
Not a framework. Not a vector database. Not a managed service. A living filesystem for an agent you talk to.
The idea is older than the implementation. What changed is the agent. Today's models are finally capable of keeping their own tooling alive. Every earlier attempt at a self-maintaining knowledge system died in the same place, where the human in the loop could no longer keep up. Myco puts the loop inside the agent. Every surface, every verb, every rule assumes the maintainer is an agent, so the human in the loop is no longer load-bearing.
You speak. The agent listens. Between your turns, Myco runs a metabolism.
hunger asks what is missing. eat absorbs whatever you point at, whether a path, a URL, or a paragraph. sense and forage scan what is already here. excrete safely removes a raw note captured by mistake, moving it to an audit tombstone instead of silently deleting. intake (v0.6.0) bulk-ingests a whole directory, composing forage + eat with strict-mode failure visibility.assimilate cooks raw notes into integrated knowledge. digest promotes a single note. sporulate concentrates integrated notes into a dispersible proposal.traverse walks the graph and reports on its connectedness. propagate publishes learning to a downstream substrate.immune runs a 47-dimension lint against the seven hard rules. senesce winds down each session cleanly.fruit drafts a three-round craft proposal, winnow gates its shape, and molt ships the contract bump.Twenty verbs, one manifest, two faces: a CLI for you to observe, an MCP server for the agent to drive. You memorize nothing. The agent drives.
integrated is a state, not an endpoint. Today's conclusion is tomorrow's raw material.Myco's own source tree is a substrate. .myco/canon.yaml at .myco/. .myco/MYCO.md as the agent entry page. .docs/primordia/ holding the three-round craft document that justifies every contract bump. The Python code under src/myco/ is the innermost ring of its own ecosystem. Not a read-only artifact someone else wrote.
So the normal install clones the source and runs pip install -e on it. The agent that uses Myco is the same agent that maintains Myco. When it needs a new lint dimension, it scaffolds one with myco ramify, writes the case with myco fruit, gates the shape with myco winnow, and ships the bump with myco molt. No fork. No waiting PR. No long-lived feature branch. ζ°ΈζθΏε.
PyPI exists for bootstrap and for library embedding. It is not the normal install path.
This clones the repository to ~/myco, runs pip install -e on it, and leaves you with a writable kernel. Then germinate a substrate inside any project:
Hook Myco into your agent host in one command:
/plugin marketplace add Battam1111/Myco, then /plugin install myco@myco.myco-install host cowork to write the MCP entry, then download myco-<ver>.zip and drag it into Claude Desktop β Settings β Plugins β Upload. Claude Desktop uploads it to your private Cowork marketplace and every subsequent session auto-installs the myco-substrate onboarding skill. (v0.7.4 hotfix: artifact extension is .zip, not .plugin β Anthropic GitHub issue #40414.)myco-install host <cursor | windsurf | zed | vscode | openclaw | claude-desktop | gemini-cli | codex-cli | goose>, or pass --all-hosts to auto-detect every host on this machine.io.github.Battam1111/myco for clients that auto-resolve it.Per-host snippets for the nine hosts with divergent schemas, Python-framework adapters (LangChain, CrewAI, DSPy, Smolagents, Agno, PraisonAI, MS Agent Framework, Claude Agent SDK), and library-embedding examples all live in INSTALL.md.
pip install 'myco[mcp]' β MCP server stack (recommended baseline).pip install 'myco[adapters]' β PDF / HTML / URL ingestion adapters (~10 MB).pip install 'myco[multimedia]' β v0.8.0: audio / image-OCR / video-frame ingestion adapters. Heavy footprint (~500 MB once PyTorch + Whisper weights download); strictly opt-in. Default install stays lean β without this extras the three multimedia adapters cleanly emit install-extras failed-stubs when forage encounters a .mp3 / .png / .mp4, so eat never silently drops a media file. Image and video OCR additionally require the system tesseract binary on PATH (install via apt-get install tesseract-ocr / brew install tesseract / Windows installer).Factual signals from GitHub, npm, and our automated checks β not a rating.
No reviews yet β be the first to share how this listing worked for you.
Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.
[](https://allmcps.com/mcp/battam1111-myco)<a href="https://allmcps.com/mcp/battam1111-myco"><img src="https://allmcps.com/api/badge/battam1111-myco?style=directory" alt="Myco on AllMCPs" /></a>