Skip to content
Agentmart
Submit a tool

Tooluniverse

2,700+ scientific tools for AI scientists: life science, research, literature, and more.

Connector by mims-harvard · GitHub account verified by the MCP Registry

IndexedApache-2.0Local (stdio)★ 1,757
Write a review

Install

Works with Claude Code, Codex, VS Code, Cursor, Windsurf, Claude Desktop. Read the repository before you give any tool access to your data.

npx agentmarket install connector/mims-harvard-tooluniverse

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

2,700+ scientific tools for AI scientists: life science, research, literature, and more.

Use cases: deep-research, data-analysis

Reviews

Write a review

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this server can do and what it asks for, from its tool list and install config. Checked Oct 25, 2026.

Runs code on install

Its package runs an install script (such as npm postinstall or setup.py code) as soon as you install it, before any tool is called. Read the script first.

Tools
4 read · 4 write · 0 delete · 4 not yet classified
Secrets
16 (API keys, tokens or passwords)
Install scripts
Runs code on install
Runs as
Python package

Tools by what they do 12

Read from the source code; the running server may list different tools.

Read 4 · Look things up without changing anything.

  • find_tools
  • get_abstract_from_patent_app_number
  • get_claims_from_patent_app_number
  • get_full_text_from_patent_app_number

Write 4 · Create, change, send or run things.

  • boltz2_docking
  • run_compass_prediction
  • run_pinnacle_ppi_retrieval
  • run_transcriptformer_embedding_retrieval

Not yet classified 4 · We could not tell from the name and description.

  • carbon_a_annotate_cds
  • compute_depmap24q2_gene_correlations
  • dynamic_tool_function
  • rfdiffusion2_design

Secrets and settings it asks for

Secrets: values you paste in that let it act as you.

  • ANTHROPIC_API_KEY
  • AZURE_OPENAI_API_KEY
  • BIOGRID_API_KEY
  • BRENDA_PASSWORD
  • DISGENET_API_KEY
  • GEMINI_API_KEY
  • HF_TOKEN
  • NCBI_API_KEY
  • NVIDIA_API_KEY
  • OMIM_API_KEY
  • ONCOKB_API_TOKEN
  • OPENAI_API_KEY
Show 4 more
  • OPENROUTER_API_KEY
  • SEMANTIC_SCHOLAR_API_KEY
  • UMLS_API_KEY
  • USPTO_API_KEY

Other settings (environment variables or headers).

  • BRENDA_EMAIL

Tools (12)

Read from the source code; the running server may differ. Descriptions are the server’s own words; hidden characters are shown as [U+…].

  • boltz2_docking

    Run the Boltz2 docking tool. Args: sequence: Protein sequence using single-letter amino acid codes. ligands: Ligand dictionaries with id and smiles fields. Returns dict: A dictionary containing the docking results with the following structure: - predicted_structure (str): The predicted protein-ligand complex structure in CIF format - structure_format (str): Format of the structure file (typically 'cif') - structure_error (str): Error message if structure file is missing - affinity_prediction (object): JSON object containing affinity predictions and r…

  • carbon_a_annotate_cds

    Annotate protein-coding genes in eukaryotic DNA with Carbon-A. Returns dict: status plus either error (a sanitized message) or data with the gene counts, the predicted coding bases, the records annotated and the genes themselves, each with coordinates, strand, exon count and the model's confidence that its structure is exactly right. A record shorter than min_record_length is refused rather than reported as having no genes, because the pipeline skips such records. A genetic code that may not fit the sequence returns the genes together with the pipel…

  • compute_depmap24q2_gene_correlations

    Run one bounded DepMap lookup without blocking the MCP event loop.

  • dynamic_tool_function

    Description is built at runtime.

  • find_toolsFind tools by descriptionread-only

    Find and search available ToolUniverse tools using AI-powered search. This tool provides the same functionality as the tools/find MCP method. Args: query: Search query describing the desired functionality categories: Optional list of categories to filter by limit: Maximum number of results to return (default: 10) use_advanced_search: Use AI-powered search if available (default: True) search_method: Specific search method - 'auto', 'llm', 'embedding', 'keyword' (default: 'auto') Returns: JSON string containing matching tools with detailed information

  • get_abstract_from_patent_app_number

    Retrieve the abstract of a patent application by its application number. Args: applicationNumberText: An 8- to 16-digit USPTO application number. Returns dict: A dictionary containing the abstract text under the 'result' key or an error message under the 'error' key if the document could not be retrieved.

  • get_claims_from_patent_app_number

    Retrieve the claims of a patent application by its application number. Args: applicationNumberText: An 8- to 16-digit USPTO application number. Returns dict: A dictionary containing the claims text under the 'result' key or an error message under the 'error' key if the document could not be retrieved.

  • get_full_text_from_patent_app_number

    Retrieve the full text of a patent application by its application number. Args: applicationNumberText: An 8- to 16-digit USPTO application number. Returns dict: A dictionary containing the full text under the 'result' key or an error message under the 'error' key if the document could not be retrieved.

  • rfdiffusion2_design

    Design protein backbones with RFdiffusion2. Returns dict: status plus either error (a sanitized message) or data with num_designs, runtime_seconds, and designs. Each design reports its name, residue and atom counts, chains, ligands, the motif residues that kept side-chain atoms, and the PDB text when requested. A missing, oversized, malformed, or non-finite design returns an error rather than a partial success.

  • run_compass_prediction

    Predict ICI response for one provider-rooted COMPASS TPM table.

  • run_pinnacle_ppi_retrieval

    Retrieve embeddings without blocking the MCP event loop.

  • run_transcriptformer_embedding_retrieval

    Retrieve embeddings without blocking the MCP event loop.

Signals

Stars
1,757
Forks
267
Open issues
9
Contributors
36
Weekly downloads
17,820
Last commit
Oct 24, 2026 (3 days ago)
Last release
Oct 13, 2026 (14 days ago)
Last package publish
Oct 13, 2026 (14 days ago)

From GitHub (fetched Oct 24, 2026) and the package registry.

Details

Registry name
io.github.mims-harvard/tooluniverse
Version
1.5.7
Auth
API key
Packages
pypi tooluniverse
License
Apache-2.0
Repository
github.com/mims-harvard/ToolUniverse
Website
aiscientist.tools
First seen
Oct 24, 2026
Last checked
Oct 24, 2026
Source
Official MCP Registrygithub-graphql
Cite this page

Tooluniverse MCP server. Agent Mart. https://agentmart.dev/connectors/mims-harvard-tooluniverse (last updated 2026-10-24)