Find and understand Synapse data without writing code.
For researchers working in Synapse, it turns dataset search, project browsing and provenance tracing into plain questions, with no code to write. Its 71 tools also reach sharing audits, wikis, challenges and schema validation, and the assistant loads only the ones a request needs.
What it is
Sage Bionetworks' own server for Synapse, hosted at mcp.synapse.org. Your assistant can search Synapse, inspect projects, folders, files and tables, and trace provenance without you writing code. It also creates, changes and deletes Synapse objects, but it never uploads or downloads file content.
What you get
- Search across public and private Synapse entities, and an exact name turned into its Synapse ID
- Metadata, annotations and the contents of a project or folder
- Provenance for an entity version: what produced it, from which inputs
- Sharing settings and your own permissions, audited across a whole project
- Wiki pages, their table of contents and revision history
- Teams, members, invitations and user profiles
- Challenge queues, submissions and scoring statuses
- JSON schemas bound to entities, and which entities fail validation
- Curation tasks and form submissions
- Entities, sharing, table columns, teams, challenge queues and schemas created, changed or deleted (metadata only)
Requirements
A Synapse account; you sign in to Synapse in your browser the first time, and what you can read or change follows your Synapse permissions. Sage warns that using it with consumer AI services that store conversations may break the Synapse Terms of Service ban on redistributing data. For sensitive or restricted data, use an enterprise AI deployment with data-residency guarantees or a self-hosted model.
Setup effort
Paste a URL, then authorize — add the endpoint to your client, then approve the OAuth consent screen