Servers Playground How we check Submit a server

Together AI MCP by usefulapi

Working · 397 ms

io.usefulapi/together-aiv1.17.0 Domain verified

Run Together AI chat, embeddings and images; manage fine-tunes, batches and endpoints.

AI & Agents Design & Creative

Source Website Registry entry

Try it

Call this server's tools from your browser. Nothing to install.

https://together-ai.usefulapi.io/mcp

Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.

Tool list from our last check. Connect live to refresh it.

together_cancel_batch

Cancel a batch job that has not finished. Together: POST /batches/{id}/cancel.

batch_idstring, required

The batch job id.

The server marks this tool as destructive. Running it may delete or overwrite data on the server or in any account you connected with your headers.

I understand, run it anyway

Together AI MCP by usefulapi MCP server | mcptry

Tools (23)

together_cancel_batch

Cancel a batch job that has not finished. Together: POST /batches/{id}/cancel.

destructive

Argument	Type	Description
batch_id*	string	The batch job id.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "batch_id"
  ],
  "properties": {
    "batch_id": {
      "type": "string",
      "minLength": 1,
      "description": "The batch job id."
    }
  }
}

together_cancel_fine_tune

Cancel a running fine-tuning job. Cannot be resumed, but a new job can continue from its last checkpoint via from_checkpoint. Together: POST /fine-tunes/{id}/cancel.

destructive

Argument	Type	Description
fine_tune_id*	string	The job id, e.g. ft-abc123.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "fine_tune_id"
  ],
  "properties": {
    "fine_tune_id": {
      "type": "string",
      "minLength": 1,
      "description": "The job id, e.g. ft-abc123."
    }
  }
}

together_chat_completion

Run a chat completion on a Together model (billed per token). Non-streaming. For a dedicated endpoint pass its `<project_slug>/<endpoint_slug>` as the model. Together: POST /chat/completions.

destructive

Argument	Type	Description
seed	integer	Seed for reproducible sampling.
stop	array	Stop sequences.
model*	string	Model name, e.g. meta-llama/Llama-3.3-70B-Instruct-Turbo.
top_k	integer
top_p	number
messages*	array	The conversation so far.
max_tokens	integer	Maximum tokens to generate.
temperature	number
reasoning_effort	string	Reasoning effort for reasoning models that support it.

together_create_batch

Start an asynchronous batch job over an uploaded JSONL input file (purpose batch-api), at a discount to real-time inference. Together: POST /batches.

destructive

Argument	Type	Description
endpoint*	string	The API each line of the input file is sent to.
model_id	string	Model to process the requests with.
priority	integer	Processing priority.
input_file_id*	string	File id of the uploaded JSONL request file.
completion_window	string	Time window for completion, e.g. 24h.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "input_file_id",
    "endpoint"
  ],
  "properties": {
    "endpoint": {
      "enum": [
        "/v1/chat/completions",
        "/v1/audio/transcriptions",
        "/v1/audio/translations"
      ],
      "type": "string",
      "description": "The API each line of the input file is sent to."
    },
    "model_id": {
      "type": "string",
      "description": "Model to process the requests with."
    },
    "priority": {
      "type": "integer",
      "maximum": 9007199254740991,
      "minimum": -9007199254740991,
      "description": "Processing priority."
    },
    "input_file_id": {
      "type": "string",
      "minLength": 1,
      "description": "File id of the uploaded JSONL request file."
    },
    "completion_window": {
      "type": "string",
      "description": "Time window for completion, e.g. 24h."
    }
  }
}

together_create_embeddings

Generate vector embeddings for one or more texts (billed per token). Together: POST /embeddings.

destructive

Argument	Type	Description
input*	any	A text, or a list of texts, to embed.
model*	string	Embedding model, e.g. BAAI/bge-large-en-v1.5.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "model",
    "input"
  ],
  "properties": {
    "input": {
      "anyOf": [
        {
          "type": "string"
        },
        {
          "type": "array",
          "items": {
            "type": "string"
          },
          "minItems": 1
        }
      ],
      "description": "A text, or a list of texts, to embed."
    },
    "model": {
      "type": "string",
      "minLength": 1,
      "description": "Embedding model, e.g. BAAI/bge-large-en-v1.5."
    }
  }
}

together_create_endpoint

Deploy a model on dedicated GPUs. The endpoint STARTS AUTOMATICALLY and bills per minute of uptime until stopped — set inactive_timeout to auto-stop it, and use together_list_hardware for valid hardware ids. Together: POST /endpoints.

destructive

Argument	Type	Description
model*	string	The model to deploy.
state	string	Initial state. Pass STOPPED to create without starting (and without billing).
hardware*	string	Hardware id, e.g. 1x_nvidia_a100_80gb_sxm.
autoscaling*	object	Replica bounds for autoscaling.
display_name	string	Human-readable name.
inactive_timeout	integer	Minutes of inactivity before auto-stop; 0 disables it.
availability_zone	string	Availability zone, e.g. us-central-4b.
disable_speculative_decoding	boolean

together_create_fine_tune

Start a fine-tuning job on an uploaded training file (billed per token processed). Stop it with together_cancel_fine_tune. Together: POST /fine-tunes.

destructive

Argument	Type	Description
model*	string	Base model to fine-tune.
suffix	string	Suffix for the fine-tuned model's name (max 64 chars).
n_evals	integer	Evaluations on the validation set during training.
n_epochs	integer	Passes over the training data.
batch_size	any	Batch size, or 'max' (the default).
warmup_ratio	number
learning_rate	number
n_checkpoints	integer	Intermediate checkpoints to save.
training_file*	string	File id of an uploaded training file (purpose fine-tune).

together_generate_image

Generate images from a prompt (billed per image/megapixel). Returns image URLs by default rather than base64, to keep responses small. Together: POST /images/generations.

destructive

Argument	Type	Description
n	integer	Number of images.
seed	integer
model*	string	Image model, e.g. black-forest-labs/FLUX.1-schnell.
steps	integer	Number of generation steps.
width	integer	Width in pixels.
height	integer	Height in pixels.
prompt*	string	What to draw.
image_url	string	Input image URL, for models that support editing.
output_format	string
guidance_scale	number

together_get_batch

Fetch one batch job: status (VALIDATING, IN_PROGRESS, COMPLETED, FAILED, EXPIRED, CANCELLED), progress, and the output_file_id / error_file_id once done. Together: GET /batches/{id}.

read-only

Argument	Type	Description
batch_id*	string	The batch job id.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "batch_id"
  ],
  "properties": {
    "batch_id": {
      "type": "string",
      "minLength": 1,
      "description": "The batch job id."
    }
  }
}

together_get_endpoint

Fetch one dedicated endpoint: state, model, hardware, autoscaling bounds and display name. Together: GET /endpoints/{endpointId}.

read-only

Argument	Type	Description
endpoint_id*	string	The endpoint id, e.g. endpoint-d23901de-....

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "endpoint_id"
  ],
  "properties": {
    "endpoint_id": {
      "type": "string",
      "minLength": 1,
      "description": "The endpoint id, e.g. endpoint-d23901de-...."
    }
  }
}

together_get_file

Fetch one file's metadata, including its processing_status and validation_report (why a fine-tune training file was rejected). Together: GET /files/{id}.

read-only

Argument	Type	Description
file_id*	string	The file id, e.g. file-abc123.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "file_id"
  ],
  "properties": {
    "file_id": {
      "type": "string",
      "minLength": 1,
      "description": "The file id, e.g. file-abc123."
    }
  }
}

together_get_fine_tune

Fetch one fine-tuning job: status, progress, hyperparameters, token counts, cost and the output model name. Together: GET /fine-tunes/{id}.

read-only

Argument	Type	Description
fine_tune_id*	string	The job id, e.g. ft-abc123.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "fine_tune_id"
  ],
  "properties": {
    "fine_tune_id": {
      "type": "string",
      "minLength": 1,
      "description": "The job id, e.g. ft-abc123."
    }
  }
}

together_list_batches

List batch inference jobs with status, progress, model and input/output/error file ids. Together: GET /batches.

read-only

No arguments.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {}
}

together_list_endpoints

List endpoints with model, owner and state (PENDING, STARTING, STARTED, STOPPING, STOPPED, ERROR). Use mine=true and type=dedicated to see what is running on your account and billing by the minute. Together: GET /endpoints.

read-only

Argument	Type	Description
mine	boolean	Only endpoints owned by the caller.
type	string	Filter by endpoint type.
usage_type	string	Filter by usage type.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {
    "mine": {
      "type": "boolean",
      "description": "Only endpoints owned by the caller."
    },
    "type": {
      "enum": [
        "dedicated",
        "serverless"
      ],
      "type": "string",
      "description": "Filter by endpoint type."
    },
    "usage_type": {
      "enum": [
        "on-demand",
        "reserved"
      ],
      "type": "string",
      "description": "Filter by usage type."
    }
  }
}

together_list_evaluations

List LLM-as-a-judge evaluation jobs (classify, score, compare) with status, parameters and results once completed. Together: GET /evaluation.

read-only

Argument	Type	Description
limit	integer	Maximum number of jobs to return.
status	string	Filter by status: pending, queued, running, completed, error, user_error.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {
    "limit": {
      "type": "integer",
      "maximum": 1000,
      "minimum": 1,
      "description": "Maximum number of jobs to return."
    },
    "status": {
      "type": "string",
      "description": "Filter by status: pending, queued, running, completed, error, user_error."
    }
  }
}

together_list_files

List uploaded data files (fine-tune, eval and batch-api inputs, plus job outputs) with size, type, purpose and validation status. Together: GET /files.

read-only

No arguments.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {}
}

together_list_fine_tune_events

List the event log of one fine-tuning job (queued, started, checkpoint saved, epoch completed, errors). The first place to look when a job failed. Together: GET /fine-tunes/{id}/events.

read-only

Argument	Type	Description
fine_tune_id*	string	The job id, e.g. ft-abc123.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "fine_tune_id"
  ],
  "properties": {
    "fine_tune_id": {
      "type": "string",
      "minLength": 1,
      "description": "The job id, e.g. ft-abc123."
    }
  }
}

together_list_fine_tunes

List fine-tuning jobs with status, base model, output model name and training settings. Together: GET /fine-tunes.

read-only

No arguments.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {}
}

together_list_hardware

List hardware configurations for dedicated endpoints with GPU type/count/memory and price in cents per minute. Pass a model to get only compatible configurations with live availability. Together: GET /hardware.

read-only

Argument	Type	Description
model	string	Only hardware compatible with this model, with availability.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {
    "model": {
      "type": "string",
      "description": "Only hardware compatible with this model, with availability."
    }
  }
}

together_list_models

List Together's models with type (chat, language, code, image, embedding, moderation, rerank), context length, organization, license and per-token pricing. Together: GET /models.

read-only

Argument	Type	Description
dedicated	boolean	Only return models that can run on dedicated endpoints.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {
    "dedicated": {
      "type": "boolean",
      "description": "Only return models that can run on dedicated endpoints."
    }
  }
}

together_start_endpoint

Start a stopped dedicated endpoint. It bills per minute of uptime until stopped. Reversible with together_stop_endpoint. Together: PATCH /endpoints/{endpointId} with state=STARTED.

destructive

Argument	Type	Description
endpoint_id*	string	The endpoint id.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "endpoint_id"
  ],
  "properties": {
    "endpoint_id": {
      "type": "string",
      "minLength": 1,
      "description": "The endpoint id."
    }
  }
}

together_stop_endpoint

Stop a running dedicated endpoint, which stops its per-minute billing. Requests to it fail until it is started again. Together: PATCH /endpoints/{endpointId} with state=STOPPED.

destructive

Argument	Type	Description
endpoint_id*	string	The endpoint id.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "endpoint_id"
  ],
  "properties": {
    "endpoint_id": {
      "type": "string",
      "minLength": 1,
      "description": "The endpoint id."
    }
  }
}

together_whoami

Identify the API key: its organization, project and project slug. The project slug forms the `<project_slug>/<endpoint_slug>` model name for dedicated-endpoint inference. A cheap way to confirm the key works. Together: GET /whoami.

read-only

No arguments.

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "properties": {}
}

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "model",
    "hardware",
    "autoscaling"
  ],
  "properties": {
    "model": {
      "type": "string",
      "minLength": 1,
      "description": "The model to deploy."
    },
    "state": {
      "enum": [
        "STARTED",
        "STOPPED"
      ],
      "type": "string",
      "description": "Initial state. Pass STOPPED to create without starting (and without billing)."
    },
    "hardware": {
      "type": "string",
      "minLength": 1,
      "description": "Hardware id, e.g. 1x_nvidia_a100_80gb_sxm."
    },
    "autoscaling": {
      "type": "object",
      "required": [
        "min_replicas",
        "max_replicas"
      ],
      "properties": {
        "max_replicas": {
          "type": "integer",
          "maximum": 9007199254740991,
          "minimum": 1,
          "description": "Maximum replicas to scale up to under load."
        },
        "min_replicas": {
          "type": "integer",
          "maximum": 9007199254740991,
          "minimum": 0,
          "description": "Replicas kept running even with no load."
        }
      },
      "description": "Replica bounds for autoscaling."
    },
    "display_name": {
      "type": "string",
      "description": "Human-readable name."
    },
    "inactive_timeout": {
      "type": "integer",
      "maximum": 9007199254740991,
      "minimum": 0,
      "description": "Minutes of inactivity before auto-stop; 0 disables it."
    },
    "availability_zone": {
      "type": "string",
      "description": "Availability zone, e.g. us-central-4b."
    },
    "disable_speculative_decoding": {
      "type": "boolean"
    }
  }
}

Input schema (JSON)

{
  "type": "object",
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "required": [
    "model",
    "training_file"
  ],
  "properties": {
    "model": {
      "type": "string",
      "minLength": 1,
      "description": "Base model to fine-tune."
    },
    "suffix": {
      "type": "string",
      "maxLength": 64,
      "description": "Suffix for the fine-tuned model's name (max 64 chars)."
    },
    "n_evals": {
      "type": "integer",
      "maximum": 9007199254740991,
      "minimum": 0,
      "description": "Evaluations on the validation set during training."
    },
    "n_epochs": {
      "type": "integer",
      "maximum": 9007199254740991,
      "minimum": 1,
      "description": "Passes over the training data."
    },
    "batch_size": {
      "anyOf": [
        {
          "type": "integer",
          "maximum": 9007199254740991,
          "minimum": 1
        },
        {
          "type": "string",
          "const": "max"
        }
      ],
      "description": "Batch size, or 'max' (the default)."
    },
    "warmup_ratio": {
      "type": "number",
      "maximum": 1,
      "minimum": 0
    },
    "learning_rate": {
      "type": "number",
      "exclusiveMinimum": 0
    },
    "n_checkpoints": {
      "type": "integer",
      "maximum": 9007199254740991,
      "minimum": 1,
      "description": "Intermediate checkpoints to save."
    },
    "training_file": {
      "type": "string",
      "minLength": 1,
      "description": "File id of an uploaded training file (purpose fine-tune)."
    },
    "training_type": {
      "oneOf": [
        {
          "type": "object",
          "required": [
            "type"
          ],
          "properties": {
            "type": {
              "type": "string",
              "const": "Full"
            }
          }
        },
        {
          "type": "object",
          "required": [
            "type",
            "lora_r",
            "lora_alpha"
          ],
          "properties": {
            "type": {
              "type": "string",
              "const": "Lora"
            },
            "lora_r": {
              "type": "integer",
              "maximum": 9007199254740991,
              "minimum": 1,
              "description": "Rank of the LoRA adapter matrices."
            },
            "lora_alpha": {
              "type": "number",
              "description": "Scaling factor applied to the LoRA adapter weights."
            },
            "lora_dropout": {
              "type": "number",
              "maximum": 1,
              "minimum": 0,
              "description": "Dropout on LoRA adapter inputs."
            },
            "lora_trainable_modules": {
              "type": "string",
              "description": "Comma-separated target modules, or all-linear for the model defaults."
            }
          }
        }
      ],
      "description": "Full fine-tune or LoRA. Together defaults to LoRA when omitted."
    },
    "max_seq_length": {
      "type": "integer",
      "maximum": 9007199254740991,
      "minimum": 1
    },
    "from_checkpoint": {
      "type": "string",
      "description": "Continue from a previous job: <job_id>, <output_model_name>, optionally with :<step>."
    },
    "training_method": {
      "oneOf": [
        {
          "type": "object",
          "required": [
            "method",
            "train_on_inputs"
          ],
          "properties": {
            "method": {
              "type": "string",
              "const": "sft"
            },
            "train_on_inputs": {
              "anyOf": [
                {
                  "type": "boolean"
                },
                {
                  "type": "string",
                  "const": "auto"
                }
              ],
              "description": "Whether prompt/user tokens contribute to the loss; 'auto' lets Together decide."
            }
          }
        },
        {
          "type": "object",
          "required": [
            "method"
          ],
          "properties": {
            "method": {
              "type": "string",
              "const": "dpo"
            },
            "dpo_beta": {
              "type": "number"
            },
            "rpo_alpha": {
              "type": "number"
            },
            "simpo_gamma": {
              "type": "number"
            },
            "dpo_reference_free": {
              "type": "boolean"
            },
            "dpo_normalize_logratios_by_length": {
              "type": "boolean"
            }
          }
        }
      ],
      "description": "Supervised fine-tuning (sft, the default) or preference tuning (dpo)."
    },
    "validation_file": {
      "type": "string",
      "description": "File id of an uploaded validation file."
    }
  }
}