Super ii MCP server

• Addresshttps://superii.site/mcpnot in the official registry: answered when asked

No entry in the official MCP registry names this server. It answered on superii.site when the Atlas asked at the address servers are usually mounted on, so it is the vendor's own. Clients and catalogues that list servers from the registry do not show it.

• Stateanswerschecked 2026-10-28
• Tools22
• Server calls itselfSuper ii public repository MCP
• Server version1.0.0
• Protocol version2025-06-18

On 2026-10-28 the server answered initialize and tools/list without a sign-in. First answer in 880 ms.

• Its tools do discover and inspect public AI models and datasets; check provenance and evidence for public Hugging Face projects, but not run open models locally on your own computer.
• Checked on 6 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: 6 so far
• What an agent can do with it2 of 3 use cases
• Answeredmeasuring6 of the 7 daily checks it takes to print a share
• Median first answer1402 ms
• Space in an agent's context≈ 3,622 tokensthe definitions of all its tools (22), which an agent carries on every turn
• Tool definitionsB · 86 of 100
• Typed results0 of 22 toolsdeclared outputSchema: results come typed, read once

• Parameters without a description: in 18 of 22

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 (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). 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_models

  Search reviewed public model repositories with metadata, size, recency, and hardware filters. Empty results are real and never filled with demos.

• search_datasets

  Search reviewed public dataset repositories with metadata, size, recency, and hardware filters. Empty results are real and never filled with demos.

• search_apps

  Search reviewed public space repositories with metadata, size, recency, and hardware filters. Empty results are real and never filled with demos.

• get_repository

  Return the reviewed public repository identity, revision, metrics, representations, compatibility, and release metadata.

• get_model_card

  Return the public model card and bounded offline model analysis for one reviewed revision.

• get_dataset_schema

  Return bounded offline dataset splits, column types, row previews, samples, statistics, and provenance.

• get_files

  List reviewed files with path, media type, byte size, SHA-256 checksum, and verified download URL.

• get_lineage

  Return typed incoming and outgoing provenance relationships plus evidence links and related public repositories.

• inspect_compatibility

  Return conservative hardware and runtime guidance, including its declared, derived, or verified confidence.

• get_tokenizer_manifest

  Return the immutable tokenizer-pack hash, engine, artifacts, context limit and verification vectors for one published text model.

• tokenize_model_text

  Encode bounded text with the exact immutable tokenizer pack for a published model. Returns IDs, pieces, offsets, count, context and pack hash.

• decode_model_tokens

  Decode bounded token IDs with the exact immutable tokenizer pack for a published model.

• get_security_status

  Return canonical evidence statuses and fail-closed boundaries without claiming unavailable controls.

• download_artifact

  Return a reviewed file URL and SHA-256 checksum. This tool does not execute code or download bytes itself.

• search_papers

  Search public paper metadata and its reviewed model, dataset, and app relationships.

• get_paper

  Return machine-readable public paper metadata and connected reviewed repositories.

• get_documentation

  Return canonical system state, public agent rules, and stable documentation URLs.

• get_agent_traces

  Return opt-in public trace metadata and hashes. Private payloads and raw tokens are never returned.

• get_repository_api

  Return stable machine-readable links, authentication boundaries, checksums, and download endpoints.

• search_products

  Read published company-provided models, robots, components and hardware. Descriptions are not product certification. Empty results are real.

• get_product

  Return the published company and product facts, resources, chosen contact channels and evidence limits. No draft or credential is returned.

• ask_product

  Retrieve published fields and their sources matching a question. This is deterministic field retrieval, not independent product validation or an inferred yes/no answer. Missing details are not_documented.