Skip to content
TokPort logo

TokPort MCP server

Which MCP servers are TokPort's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.

read 2026-10-29

← Back to TokPort

Whose server it is
  • Registry nameio.tokport/tokportverified in the MCP registry

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

When we asked itchecked 2026-10-29
  • Stateanswerschecked 2026-10-29
  • Tools8
  • Server calls itselftokport-mcp
  • Server version1.0.0
  • Protocol version2025-06-18

On 2026-10-29 the server answered initialize and tools/list without a sign-in. First answer in 130 ms.

Should an agent connect itConnect it, with care
  • Its tools do every use case we checked: create short video clips billed per second; switch between image models by changing the model name.

Once connected

  • For "create short video clips billed per second", the likely first call is search_models.
  • For "switch between image models by changing the model name", the likely first call is search_models.
  • cancel_job change data without a sign-in: anyone with the address can call them.
  • Parameters are untyped or undescribed in 3 of 8 tools: check arguments before calling.

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.

Measured dailydaily checks: 7, the first on 2026-10-29
  • INITE MCP score76 of 100answers 100 · does the job 100 · definitions 77 · context 100 · connects 100 · lowered by the warning signs below
  • What an agent can do with it2 of 2 use cases
  • Answered100% of days7 of 7 daily checks
  • Median first answer269 ms
  • Space in an agent's context≈ 813 tokensthe definitions of all its tools (8), which an agent carries on every turn
  • Tool definitionsB · 77 of 100
  • Typed results0 of 8 toolsdeclared outputSchema: results come typed, read once
  • Parameters without a description: in 3 of 8
  • Not saying whether it only reads or can destroy: 8 of 8

What an agent can do with it

  • not sure · generate images with an existing OpenAI SDK client
  • can do · create short video clips billed per second
  • can do · switch between image models by changing the model name

Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?

Worth a look

  • Tools that change data, on a server that answers without a sign-in: cancel_job

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 →

Tools it offers8 tools · read 2026-10-29
  • search_models

    Search the model catalog. Returns ids, modality, tasks, pricing and links. Use before calling run_model or submit_job to pick a model id.

  • get_model_schema

    Input parameters (name, type, allowed values, defaults) and output shape for every task a model supports.

  • get_pricing

    Live price per image or per video clip for one model or all models.

  • run_model

    Run a model synchronously with the caller's API key. Images return URLs; videos are queued and this waits up to 60 s before returning the queue status instead.

  • submit_job

    Queue a request and return its request_id immediately. Follow with check_job / get_job_result.

  • check_job

    Status of a queued request (IN_QUEUE / IN_PROGRESS / COMPLETED / FAILED / CANCELLED) with position, logs and metrics.

  • get_job_result

    Result of a completed request; waits up to `timeout` seconds (max 60) when still running.

  • cancel_job

    Cancel a request while it is still queued.

How to connect

Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:

{
  "mcpServers": {
    "tokport": {
      "type": "http",
      "url": "https://tokport.io/api/mcp"
    }
  }
}

Add it

Claude Code, in a terminal

claude mcp add --transport http tokport https://tokport.io/api/mcp

Add to CursorAdd to VS Code

claude.ai, ChatGPT and other clients: add https://tokport.io/api/mcp as a custom connector in the settings.

Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.