TokenLab MCP Server
@hedging8563
基本信息
关于 TokenLab MCP Server
TokenLab MCP integration for model discovery, pricing, typed LLM requests, multimodal generation, async tasks, files and developer API workflows.
TokenLab MCP Server 配置
使用下面的配置,将此服务器添加到你的 MCP 客户端。
{
"mcpServers": {
"tokenlab": {
"command": "npx",
"args": [
"-y",
"@tokenlabai/mcp-server"
]
}
}
}TokenLab MCP Server 工具
87Cancels an Anthropic Message Batch when cancellation is still available for that batch.
Cancel batch
Cancels a queued asynchronous task when cancellation is supported for the selected task. Cancellation currently supports queued Seedance video tasks (`seedance-2.0` and `seedance-2.0-fast`) while they are still waiting to run. Tasks that are already processing, completed, failed, expired, or unsupported are not cancelled.
Compacts Responses input and returns the native `response.compaction` contract. The request model must be available on a same-protocol Responses path.
Counts input tokens for an Anthropic-compatible messages request without creating a model response. Authentication is required; the operation is not billed. Verified upstream counting is preferred. Upstream availability failures return a local estimate marked by X-TokenLab-Count-Tokens-Source: local_estimate. Platform authentication, permission and input errors remain errors. Bound files unreadable locally require upstream counting.
Creates a 3D model generation task using Tripo3D. Returns a task ID for polling.
Creates a model response using Anthropic's native request format. Supports Claude models with vision and streaming. **Authentication:** Use `x-api-key` header or `Authorization: Bearer` header.
Creates a native Anthropic Message Batch. Requests use the same Anthropic message parameters as /v1/messages.
Creates an OpenAI-compatible batch job.
Create an OpenAI-compatible chat completion. For inline image_url data URLs, declare the byte-accurate image MIME type. The MCP boundary corrects recognized PNG, JPEG, WebP, and GIF payloads declared as application/octet-stream and rejects unrecognized generic binary image payloads before sending a billable request.
Creates an embedding vector representing the input text.
Generates content using the Gemini GenerateContent request shape. The model details must advertise Gemini requests and the model must be available. ProtoJSON lowerCamelCase and original proto snake_case field names are preserved. Unknown fields are accepted on a best-effort basis; model-specific support is not guaranteed. **Authentication:** Use `?key=` query parameter, `x-goog-api-key` header, or `Authorization: Bearer` header.
Creates an image given a prompt.
Creates an image given a prompt.
Generates image variations from a multipart image upload.
Creates embeddings for multimodal input items. Text input is generally available; image input may require feature enablement.
Creates a music generation task using Suno. Returns a task ID for polling.
Creates a response using the OpenAI Responses API request shape. The model details must advertise the Responses request format and the model must be available. Unknown fields are accepted on a best-effort basis; model-specific support is not guaranteed.
Uploads one Seedance material asset URL into a TokenLab material group. Image materials must use supported dimensions; an invalid source is rejected with 400 seedance_material_source_invalid instead of remaining in PROCESSING. Real-person assets require the group_id mapped from Result.GroupId returned by GetVisualValidateResult.
Creates a TokenLab Seedance virtual-avatar material group. Real-person material groups are created by GetVisualValidateResult after the H5 validation succeeds.
Generates audio from the input text (Text-to-Speech).
Creates an asynchronous video generation task. The response returns a canonical task ID and usually a preferred `poll_url`; clients should poll `poll_url` first, or use `/v1/tasks/{id}` as the fixed status endpoint.
Create a real-person visual validation session through the Volc-compatible Action API. Returns the official ResponseMetadata and Result shape with BytedToken, H5Link, and CallbackURL.
Creates a Seedance 2.0 video task with the official Volc REST body and response shape. Authentication uses a TokenLab Bearer API key. TokenLab additionally accepts an optional Idempotency-Key header so a client can recover the same task ID after a lost create response.
Create a workspace webhook; save the returned signing secret. Requires TOKENLAB_MANAGEMENT_TOKEN (mt-...) for the workspace; sk-... keys are not accepted.
Creates a World Labs Marble world generation task for TokenLab-supported Marble models. Returns a task ID for polling. TokenLab supports native world_prompt media_asset references when the media asset was prepared through TokenLab.
Deletes an Anthropic Message Batch when deletion is supported for that batch.
Deletes a batch file or a bound image/reference file. When the request includes anthropic-beta: files-api-2025-04-14, the route uses Anthropic Files API mode for the bound workspace.
Deletes a stored Responses API response. This operation is deletion, not cancellation of an in-progress response.
Deletes a Seedance material asset owned by the authenticated organization.
Deletes a TokenLab Seedance material group and its material assets. The operation is irreversible when accepted.
Cancel Seedance task (Volc-compatible)
Delete a webhook and its delivery history. Requires TOKENLAB_MANAGEMENT_TOKEN (mt-...) for the workspace; sk-... keys are not accepted.
Deletes a World Labs world generated through TokenLab. This permanently deletes the world and associated assets when World Labs accepts the deletion. TokenLab keeps task status records for usage history.
Dispatches Volc-style Action requests. Use Version=2024-01-01 or omit Version.
Dispatches Volc-style Action requests with TokenLab Bearer authentication. Real-person validation and material asset Actions use Version=2024-01-01 and are POST-only. The same POST contract is also available at the API root for endpoint-only migration from Volcengine.
Edits an image using multipart image uploads, JSON image URLs, or the official JSON images array for supported GPT Image models.
Edits an image using multipart image uploads, JSON image URLs, or the official JSON images array for supported GPT Image models.
Evaluate typed Noul probability, Choice, and Score questions against shared state using the native System One endpoint. Discover a decision model with list_models/get_model first. Returns synchronous structured decisions, not chat text. A decision does not authorize an external action.
Creates multiple embeddings using the native Gemini BatchEmbedContents contract.
Counts tokens using the native Gemini CountTokens contract. Unknown fields and either official ProtoJSON field-name spelling are forwarded best-effort.
Creates an embedding using the native Gemini EmbedContent contract.
Retrieves the status and result of a 3D generation task.
Retrieves model metadata through the Gemini-compatible v1beta surface.
Compatibility alias for polling public async image task IDs such as `ldtask_...`. Prefer the `poll_url` returned by create responses, or use `/v1/tasks/{id}` as the canonical fixed status endpoint. This route uses the same authorization, ownership checks, billing headers, terminal snapshots, and response shape as the image-specific status handler.
Retrieves current model details, including supported operations, request parameters, request endpoints, selector limits, and pricing when available. Non-chat integrations should read these current model-specific request details before creating a request.
Retrieves pricing-only detail for one model. Use this endpoint for price explanation, not for non-chat request construction.
Retrieves the status and result of a music generation task.
Returns the public pricing surface for active models, with optional provider and tag filters.
Retrieves one TokenLab Seedance material asset and its asynchronous ingestion status.
Retrieves one TokenLab Seedance material group.
Retrieves the status and result of an asynchronous generation task. Prefer the `poll_url` returned by the create response; `/v1/tasks/{id}` is the canonical fixed status endpoint for video, music, and 3D jobs.
Retrieves the status and result of a video generation task. New clients should prefer the `poll_url` returned by create responses, or use `/v1/tasks/{id}` as the canonical fixed status endpoint. The `/v1/videos/generations/{id}` path remains for backward compatibility.
Get the mapped real-person material GroupId through the Volc-compatible Action API after the H5 validation succeeds. The operation returns the official ResponseMetadata and Result shape.
Get Seedance task (Volc-compatible)
Read a webhook without revealing its secret. Requires TOKENLAB_MANAGEMENT_TOKEN (mt-...) for the workspace; sk-... keys are not accepted.
Retrieves a World Labs world generated through TokenLab. The world ID must be owned by the authenticated organization.
Retrieves metadata for a World Labs media asset prepared through TokenLab. The media asset must belong to the authenticated organization.
Retrieves the status and result of a World Labs Marble world generation task.
Lists native Anthropic Message Batches for the authenticated organization.
List batches
List files
Lists enabled models through the Gemini-compatible v1beta model discovery surface.
Lists the currently available models. Use `view=compact` for model selection and `view=full` for the existing OpenAI-compatible discovery shape. Native API calls default to `full`; integrations may choose `compact` as their default. Non-chat recommendations are available through `recommended_for`.
Lists Seedance material groups owned by the authenticated organization. Real-person groups are returned after GetVisualValidateResult succeeds.
Lists Seedance material assets owned by the authenticated organization. Use library_type, group_id, and status filters to find assets for a generation workflow.
Lists recent Seedance compatibility tasks for the authenticated organization.
List delivery outcomes, HTTP statuses and cumulative attempts. Requires TOKENLAB_MANAGEMENT_TOKEN (mt-...) for the workspace; sk-... keys are not accepted.
List workspace webhooks. Requires TOKENLAB_MANAGEMENT_TOKEN (mt-...) for the workspace; sk-... keys are not accepted.
Lists World Labs worlds created through TokenLab for the authenticated organization. Results only include resources created through TokenLab.
Creates a World Labs media asset record and returns a signed upload URL. Use the returned media_asset_id from native world_prompt objects with source=media_asset. Media assets can only be used by the organization that created them.
Reranks documents by relevance to a query using semantic similarity.
Retrieves the status and metadata for a native Anthropic Message Batch.
Downloads JSONL results for a completed native Anthropic Message Batch.
Retrieve batch
Retrieve file
Retrieve file content
Retrieves a stored Responses API response by TokenLab response id. Ownership checks return 404 when the response is missing or unavailable to the current API key.
Replace the signing secret and return the new secret once. In-flight deliveries may use the previous secret. Requires TOKENLAB_MANAGEMENT_TOKEN (mt-...) for the workspace; sk-... keys are not accepted.
Send webhook.test; inspect outcome to confirm receipt. Requires TOKENLAB_MANAGEMENT_TOKEN (mt-...) for the workspace; sk-... keys are not accepted.
Transcribes audio into text (Speech-to-Text).
Transcribes audio and translates the result to English.
Translates text into a target language using the current translation request format.
Updates a Seedance material asset owned by the authenticated organization.
Updates display metadata for a TokenLab Seedance material group.
Update URL, events, description or active state. Resuming resets failures. Requires TOKENLAB_MANAGEMENT_TOKEN (mt-...) for the workspace; sk-... keys are not accepted.
Uploads a TokenLab/OpenAI-compatible batch or image file. With anthropic-beta: files-api-2025-04-14, uploads an Anthropic Files resource without a purpose field and returns Anthropic file metadata.
概览
What is TokenLab MCP Server?
TokenLab MCP integration for model discovery, pricing, typed LLM requests, multimodal generation, async tasks, files and developer API workflows.
How to use TokenLab MCP Server?
Follow the linked project README for installation and MCP client configuration.
Tools
See the project documentation for its current tool inventory.
Documentation
Tool profiles
Version 0.6.35 defaults to the core profile (32 tools). The full profile exposes 89 tools; the manifest inventory below includes optional profile tools. See the linked documentation for profile configuration.
AI 与智能体 分类下的更多 MCP 服务器
scvd.store
seancrecordscvd.store is an evidence observatory for agentic commerce. Independent signed observation of what other people's endpoints, artifacts and payments actually did among other things. USDC on Base, Arbitrum One, World Chain
fhirHydrant
faulkjOpen-source Node.js FHIR MCP server with SMART Backend Services, metadata-aware search/CRUD tools, compact responses, FHIRPath filtering, safe pagination, audit events, and terminology lookup.
Cartesia MCP
cartesia-aiCartesia MCP is a remote MCP server for Cartesia (https://cartesia.ai) available at https://mcp.cartesia.ai.
Legion MCP
faulkjMCP-native LLM councils for debates, juries, blind panels, voting, refinement, and custom multi-model deliberation.
Heterogent
Hire AI agents and lock USDC in trustless on-chain escrow on Solana all from your AI assistant. Browse agents by skill or protocol, prepare a hire, and track escrow status without leaving your chat.
评论