Tomesphere MCP server

• Registry namecom.tomesphere/tomesphereverified in the MCP registry

The registry name sits under tomesphere.com. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.

• Stateanswerschecked 2026-10-21
• Tools9
• Server calls itselftomesphere-mcp
• Server version0.1.0
• Protocol version2025-03-26

On 2026-10-21 the server answered initialize and tools/list without a sign-in. First answer in 267 ms.

• Whether its tools do the product's job has not been judged yet.
• Checked on 5 of the 7 days it takes to judge reliability.

Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.

• INITE MCP scoreafter a weeka score needs 7 daily checks: 5 so far
• Answeredmeasuring5 of the 7 daily checks it takes to print a share
• Median first answer277 ms
• Space in an agent's context≈ 1,092 tokensthe definitions of all its tools (9), which an agent carries on every turn
• Tool definitionsA · 100 of 100
• Typed results0 of 9 toolsdeclared outputSchema: results come typed, read once

• Every tool passes all six checks.

A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (30%), doing the product's job (25%), definitions (20%), context (10%) and use (15%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →

• search_papers

  Search 8.5 million academic papers (arXiv + biomedical: PMC / bioRxiv / medRxiv, all disciplines) by topic, keyword, author, or linked entity (gene / protein / disease). Each hit returns id, title, TLDR, type, and links. Use to find…

• get_paper

  Fetch a paper's full metadata: title, authors, year, abstract, TLDR (LLM-generated), key findings, citation count, GitHub repos, HuggingFace models/datasets, videos, peer reviews, and links. Accepts an arXiv ID (e.g. '2401.12345' or…

• similar_papers

  Find papers semantically similar to a given paper using SPECTER2 embeddings (trained on scientific-citation triplets). Returns nearest neighbors with TLDR. Use when the user wants 'papers like X' or 'what's adjacent to this work'.

• citations

  Get papers that cite the given paper (who refers to this work). Use when the user asks 'who cites X', 'what's the impact', or wants follow-up work. Note: 2024+ citation coverage is sparse; indexing in progress.

• references

  Get the papers that this paper cites (its bibliography). Use when the user asks 'what does X cite' or 'show me the references'.

• get_full_text

  Fetch a paper's full body text as Markdown (methods, results, protocols, inline tables) — use for deep questions the abstract can't answer. Accepts an arXiv ID (2401.12345), a PMC ID (PMC5339222), or a bioRxiv/medRxiv DOI (10.1101/…).

• get_figures

  Get a paper's real figure images — URLs, labels, and captions. Most biomedical papers have figures; arXiv papers often don't. Accepts an arXiv ID, PMC ID, or bioRxiv/medRxiv DOI.

• get_entities

  Get the biomedical entities linked to a paper: genes, proteins, chemicals, diseases, species, mutations, cell lines, and clinical-trial (NCT) IDs. Accepts an arXiv ID, PMC ID, or bioRxiv/medRxiv DOI.

• get_structure

  Resolve a gene/protein name (e.g. 'TP53', 'CD44') or UniProt accession to its 3D structure — returns the UniProt accession + AlphaFold model URL (and PDB when available). Great for a gene named in a paper's entities.