The ordinary create-and-update tools are the smaller half. The interesting ones aggregate: a project's context in a single call rather than four, related content for any entity, and a natural-language question answered across projects, tasks and documents together. The conversation tools close the loop — a planning discussion can be saved against its project, then mined later for the action items nobody wrote down. Give the API key only the permissions the workflow needs; it is the boundary on everything here.
A full client for the Helios-9 project management service. Everything the workspace holds — the project hierarchy, the work items, the documents, past AI conversations — is readable and writable, with context-shaped calls that aggregate rather than return one record at a time.
- Projects created, read, updated, archived, duplicated and updated in bulk, with a timeline of milestones and key events
- Initiatives as a layer above projects, created and updated, with their own context and insight calls and documents linked to or unlinked from them
- Tasks filtered by project, initiative, status or assignee, created and updated singly or in bulk, with dependencies between tasks recorded and read, and multi-task workflows created and tracked to completion
- Markdown documents with YAML frontmatter, searched by content with ranking, read with their links and references, analysed, generated from a template, worked on with collaborators, and operated on in bulk
- AI conversations saved against a project and retrieved later, analysed for themes and patterns, summarised in several forms, and mined for action items that can become tasks
- Context aggregation on demand: a project's statistics and recent activity, a workspace overview, related content for any entity, and a natural-language query answered across projects, tasks and documents at once
- Automation: rules created against events, listed globally or per project, executed manually for testing, and reported on for how they are performing
- Search across every entity type with ranking, semantic similarity search over embeddings, autocomplete suggestions, and analytics on the searches themselves
- Analytics on projects, team productivity and overall workspace health, with custom reports specified by metric
- A natural-language project description turned into structured entities, with an analyse mode that shows the plan before anything is created
- An environment check for when the connection itself is what is wrong
A Helios-9 account and an API key generated in its settings — it is shown once, and its permissions decide what the connection can reach across projects, tasks, documents and analytics. HELIOS_API_URL points at the service and HELIOS_API_KEY carries the key, or the key can be passed on the launch command. Node.js 16 or newer; it runs as npx -y helios9-mcp-server@latest. All state lives in the Helios-9 workspace rather than locally, so this is a client for an account you already have rather than something that stands alone.
One command plus a key — npx -y helios9-mcp-server@latest --api-key YOUR_HELIOS9_API_KEY, then supply credentials