The nervous system for AI-assisted software teams
Myco is the nervous system for AI-assisted software teams. It captures what happens across your coding sessions, turns raw memory into durable project knowledge, and routes the right context back to every agent and teammate working in the Grove.
Myco works automatically alongside the coding agents, subagents, and agent teams you already use. It does not replace their reasoning, memory, tools, or native workflows, and it does not lock you into one ecosystem.
Instead, Myco gives them an enhanced nervous system: richer senses, durable project knowledge, contextual recall of decisions and pitfalls, semantic codebase awareness, and learned workflows that compound across sessions and teammates.
Named after mycorrhizal networks, Myco is not another coding agent. It is the shared knowledge, signal layer, and coordination system beneath the agents you already use.
That distinction matters. Basic memory systems preserve snapshots: what happened, when it happened, and what was said. Myco treats memory as living project material. Some observations become obsolete as the code changes; repeated observations become wisdom; hard-won gotchas become reusable guidance. The goal is to preserve the tribal knowledge that used to live in the heads of long-running human teams, so each new coding agent can inherit accumulated context instead of starting cold.
That knowledge also evolves. Myco is not an ever-growing static archive. It keeps refining what the project knows, what the team has learned, and which workflows are still true. Because Myco lives with the project rather than inside any single coding agent or hosted harness, every connected agent can participate in that evolution without forcing your team into one AI ecosystem.
- Captures the work — sessions, prompts, tool use, plans, screenshots, decisions, trade-offs, and gotchas.
- Builds project knowledge — durable spores, wisdom, summaries, digests, graph links, and searchable context.
- Evolves what matters — stale observations can be superseded, repeated lessons become wisdom, and workflows stay current as the code changes.
- Routes context to agents — session briefings, relevant spores, and Canopy file anatomy surface when agents need them.
- Maps the codebase — Canopy gives agents semantic awareness of files before they open them.
- Learns workflows — repeated procedures become reviewed skills that every connected agent can follow.
- Shares across teams — an optional Team Host connection routes a project's knowledge to a teammate's Myco directly and exposes it to non-member agents through a secure read-only MCP endpoint, no cloud account required.
macOS is the primary supported platform. Linux and Windows are in beta. On Windows, only x64 is supported — Windows on ARM (which runs the x64 build under emulation) is not supported.
curl -fsSL https://myco.sh/install.sh | sh
myco openOn Windows x64 (PowerShell):
irm https://myco.sh/install.ps1 | iexMyco is a self-contained native binary — no Node runtime required. The installer downloads the binary to ~/.myco/bin (%LOCALAPPDATA%\Myco\bin on Windows), starts the managed local service, and connects supported coding agents. Then myco open launches the dashboard. Open any git project in a supported coding agent and Myco auto-registers it into your default Grove when the agent starts working there.
You can also open the dashboard directly at http://localhost:20915/. If your local install reports a different service URL, myco open will open the right one.
Already have Node? npm install -g @goondocks/myco also works — it's a thin bootstrap that converges to the same native binary.
Provider configuration is optional at install time. Capture and full-text search work immediately; spores, digests, semantic search, Canopy summaries, and skill lifecycle features become active after you configure intelligence and embedding providers in the dashboard.
See Quickstart for setup details and platform notes.
Myco keeps itself up to date automatically — the local service self-updates from the release channel in the background while it's idle. You can also trigger an upgrade from the Upgrade section of the dashboard's Settings page.
No npm update is needed. (A myco upgrade CLI exists for advanced or scripted use, with --channel stable|beta, but the automatic and dashboard paths are the normal way to stay current.) Upgrading from an older per-project install archives legacy Myco-owned files the next time Myco starts. See Upgrading Myco.
Team Host operator commands (myco host, myco join, myco attach) are part of the main binary and upgrade with it — no separate package. One optional operator package remains for infrastructure administration (an npm/Node tool that requires Node 22+):
@goondocks/myco-collective— deploy and manage a Myco Collective
Symbionts connect Myco to the agents you already use. They capture session starts, prompts, tool calls, stops, transcripts, and attachments into the Grove vault without replacing the agent's own memory or workflow.
The agent harness runs Myco's background intelligence work. It reads captured sessions, extracts spores (decisions, gotchas, discoveries, trade-offs, fixes), generates titles and summaries, links entities into a knowledge graph, and refreshes digest extracts.
When Myco finds 3+ semantically similar spores, it synthesizes them into a wisdom spore — a higher-order observation that captures the pattern across sessions. Individual observations become institutional knowledge.
As the project changes, that knowledge keeps moving. New sessions can reinforce a pattern, replace an outdated assumption, or expose a workflow that should become shared practice.
Every task can use a different LLM provider. Run title generation on a fast local model via Ollama, extraction on Claude, and consolidation on a larger local model via LM Studio. Configure providers from the dashboard.
See the Intelligence Pipeline docs for the task catalog, provider configuration, and scheduling.
The digest synthesizes accumulated knowledge into tiered extracts — pre-computed context at different depths:
| Tier | Purpose |
|---|---|
| 1,500 tokens | Executive briefing — what this project is, what's active, what to avoid |
| 5,000 tokens | Deep onboarding — trade-offs, patterns, team dynamics |
| 10,000 tokens | Institutional knowledge — full thread history and design tensions |
Extracts refresh in the background as new knowledge arrives. When the project goes quiet, refresh slows; new sessions wake it back up.
Every record is indexed for both keyword search and semantic similarity. Use Ollama locally for embeddings, or OpenRouter / OpenAI in the cloud. The index is fully rebuildable from the database.
Myco keeps a fresh per-file index of your project — exports, imports, top comment, optional one-line summary — and hands the agent that anatomy before it opens a file. A single myco_cortex call with op: "canopy_map" returns the project's architectural overview, so a new agent can orient in one tool call instead of a dozen searches. Manage it from the dashboard's Cortex tab. See the Canopy docs.
Myco routes project context to agents automatically:
- Session start — a project briefing gives the agent pre-computed project understanding before it asks a single question.
- Per-prompt — relevant spores are retrieved after user prompts, providing targeted context for the task at hand.
- Pre-read — Canopy file anatomy appears before reads so agents can choose the right file faster.
Agents can still use their own memory and tools. Myco adds shared project context without taking those systems over.
A local web dashboard provides configuration and operations management. Manage Groves and projects, configure providers, approve skill candidates, trigger intelligence and digest cycles, monitor service health, and view live logs.
Use the Grove Management guide to decide when to create additional Groves, move or archive projects, and toggle per-project capabilities such as Cortex, Canopy, Skills, and Vault Evolution.
Myco integrates with coding agents through symbionts. Each symbiont connects Myco to an agent's native context, tools, skills, and permissions while preserving that agent's own workflow.
Supported symbionts include Claude Code, Cursor, Codex, Cline, Copilot, Antigravity, Devin Desktop, OpenCode, and Pi. See the Symbiont docs for agent-specific details.
Share knowledge across machines and teammates with no cloud account. One teammate's Myco becomes the team's shared home; everyone else joins it directly from the dashboard:
myco host enable --server-url https://your-host:8080 # on the host machine
myco join <host> --key <one-time-key> --server-url <url> --overlay-address <address> # on a member's machineBoth steps are also available from the dashboard's Team page — enabling a host prints a ready-to-paste join command, and joining is a form with the same four fields. Once joined, connect the projects you want the team to share (Team page, or myco attach [path] --host <id>). From then on, each connected project's knowledge reaches the host directly: new spores, session summaries, plans, and graph edges become available to the team in the background, and search returns both local and team knowledge.
Local Grove databases remain the source of truth. Joined machines and the host talk over a direct, encrypted overlay connection, and each record carries a machine identity for attribution.
See the Team Host docs for the full guide, including running a team server and the external read-only MCP endpoint for agents that aren't Myco members.
Search across projects and manage shared settings by connecting multiple team workers to one Myco Collective.
Install it only if you want the cross-project admin layer:
npm install -g @goondocks/myco-collective
myco-collective installThe Collective gives you a worker-hosted admin UI for connected projects, shared settings, and cross-project search. See the Collective guide.
A Team Host can also expose a read-only, six-tool MCP endpoint over a secure public URL — for agents that aren't Myco members, like managed agents and automation platforms. Connect any tool that speaks MCP and it gets read access to the same team knowledge your local agents already have. See the Team Host docs for the tool reference and setup.
Memory is table stakes. Myco goes further: it turns accumulated project knowledge into repeatable workflows that every agent follows. The intelligence pipeline identifies procedural patterns across sessions — debugging the build, adding API routes, configuring providers, resolving common gotchas — and surfaces them as candidates. You approve what becomes canon, and Myco generates validated SKILL.md files under .agents/skills/, symlinked into every agent's native skills directory.
Skills evolve as your code does. When a pattern is abandoned, a new gotcha is discovered, or a workflow shifts, Myco refreshes affected skills — preserving what's still accurate, incorporating what's new, and splitting skills that have grown too broad. See the Skills docs for the full lifecycle.
Backups and restores are Grove-scoped. Local backups run automatically during idle periods, and destructive project deletion creates a fresh Grove backup first when backups are enabled. Configure a custom backup directory in the Backup & Restore section of the Settings page. Restore preserves existing records and avoids importing duplicates.
myco doctorVerifies your local Myco install, Grove data, provider setup, connected agents, service status, and dashboard access. Use --fix to auto-repair fixable issues.
myco remove # removes Myco's contributions from every agent's global config
myco remove --purge # also removes ~/.myco/ itselfRemoval preserves any user-pre-existing keys in agent config files (e.g. a Codex [features].hooks entry you added yourself stays put).
Contributions welcome. See the Contributing Guide for development setup, and the Lifecycle docs for architecture details. Please open an issue to discuss before submitting a PR.