client.datasets.create()

import { createClient } from "@toloka-platform/sdk";

const client = createClient({
  baseUrl: "https://platform.toloka.ai",
  apiKey: process.env.TOLOKA_API_KEY!,
});

const dataset = await client.datasets.create({
  projectId: "pp.019a0000-0000-7000-8000-000000000001",
  name: "training_examples"
});
console.log(dataset);

{
  "id": "ds.019a0000-0000-7000-8000-000000000001",
  "projectId": "pp.019a0000-0000-7000-8000-000000000001",
  "name": "<string>",
  "hidden": false,
  "profile": null,
  "createdAt": "2026-01-01T00:00:00Z"
}

import { createClient } from "@toloka-platform/sdk";

const client = createClient({
  baseUrl: "https://platform.toloka.ai",
  apiKey: process.env.TOLOKA_API_KEY!,
});

const dataset = await client.datasets.create({
  projectId: "pp.019a0000-0000-7000-8000-000000000001",
  name: "training_examples"
});
console.log(dataset);

{
  "id": "ds.019a0000-0000-7000-8000-000000000001",
  "projectId": "pp.019a0000-0000-7000-8000-000000000001",
  "name": "<string>",
  "hidden": false,
  "profile": null,
  "createdAt": "2026-01-01T00:00:00Z"
}

Send an optional name and profile. A profile creates typed fields and mappings that validate file uploads. Without a profile, uploaded rows are stored in the “uploaded” field. Returns the dataset.

Signature

Method signature on client.datasets:

create(args: {
  projectId: string;
  name?: string;
  profile?: "instruction" | "chat-sft";
}): Promise<Dataset>;

Parameters

string

required

Project ID with prefix pp..

string

Name of the dataset; defaults to New Dataset.

"instruction" | "chat-sft"

Create typed fields and upload mappings with “chat-sft” (messages, tools, split, assistant, source_ref) or “instruction” (prompt, completion, split). Without a profile, uploaded rows are stored in the “uploaded” field.

Returns

Promise<Dataset> A dataset within a project.

DatasetId

required

Dataset ID with prefix ds..

ProjectId

required

Project ID with prefix pp..

string

required

Name of the dataset.

boolean

required

Whether the dataset is hidden from ordinary dataset lists.

DatasetProfileStamp | null

required

Profile used to create the dataset fields, or null for a dataset without a profile.

Show Fields

"chat-sft" | "instruction"

required

Profile name: chat-sft for chat training examples or instruction for prompt and completion pairs.

number

required

Version of the profile definition used when the dataset was created.

string

required

Time the dataset was created, as an ISO 8601 timestamp.

Errors

HTTP errors throw PotatoApiError. The endpoint documents these error responses:

Status	Description
404	Project not found
429	Rate limit exceeded

HTTP endpoint: POST /projects/{projectId}/datasets.

import { createClient } from "@toloka-platform/sdk";

const client = createClient({
  baseUrl: "https://platform.toloka.ai",
  apiKey: process.env.TOLOKA_API_KEY!,
});

const dataset = await client.datasets.create({
  projectId: "pp.019a0000-0000-7000-8000-000000000001",
  name: "training_examples"
});
console.log(dataset);

{
  "id": "ds.019a0000-0000-7000-8000-000000000001",
  "projectId": "pp.019a0000-0000-7000-8000-000000000001",
  "name": "<string>",
  "hidden": false,
  "profile": null,
  "createdAt": "2026-01-01T00:00:00Z"
}

Reference

TypeScript SDK

CLI

HTTP API

Signature

Parameters

Returns

Errors