@toloka-platform/sdk 0.1.0-preview.9 → 0.2.0-preview.1

package/README.md

@@ -2,7 +2,7 @@
 
 Build and operate Toloka pipeline projects as TypeScript source.
 
-> **Preview.** This SDK is not stable. The authoring API, the CLI surface, and the generated source layout can change without a major version bump. Pin an exact version. Do not depend on this in production yet. Read [limitations](./docs/limitations.md).
+> **Beta.** This SDK is not stable. The authoring API, the CLI surface, and the generated source layout can change without a major version bump. Pin an exact version. Do not depend on this in production yet. Read [limitations](./docs/limitations.md).
 
 A pipeline project is a graph of nodes that moves items from a dataset through deterministic code and human work and back into the dataset. This package lets you write that project as TypeScript, keep it in version control, and reconcile it onto the platform with one command.
 
@@ -43,16 +43,17 @@ toloka init
 
 ## Install
 
-Install the bytecode-cached CLI and add the Preview SDK to the project:
+Install the bytecode-cached CLI and add the Beta SDK to the project:
 
 ```sh
 curl -fsSL https://platform.toloka.ai/cli/install.sh | sh
-bun add @toloka-platform/sdk@preview
+bun add @toloka-platform/sdk@latest
 ```
 
 The installer requires Bun 1.4.0, installs `@toloka-platform/sdk` globally, and exposes `toloka` through Bun's global binary directory. The project install makes SDK imports resolvable.
 
-For a guided first run, start a fresh coding-agent session and ask it to follow the [agent onboarding guide](https://platform.toloka.ai/onboarding.md).
+The supported package channel is `latest`; version `0.2.0-preview.1` remains Beta.
+For a guided first run, start a fresh coding-agent session and ask it to follow the [data-labeling agent guide](https://platform.toloka.ai/agents/data-labeling.md). Fine-tuning and gisting have their own guides at `/agents/fine-tuning.md` and `/agents/gisting.md`.
 
 ## Documentation
 
@@ -65,7 +66,7 @@ For a guided first run, start a fresh coding-agent session and ask it to follow
 | [Datasets](./docs/datasets.md)               | columns, immutable schemas, loading rows, what plan refuses                  |
 | [CLI](./docs/cli.md)                         | every command, flag, exit code, and the reconcile loop                       |
 | [Troubleshooting](./docs/troubleshooting.md) | the failures you will actually hit, and what to do about each                |
-| [Limitations](./docs/limitations.md)         | Preview status and what this SDK does not do yet                             |
+| [Limitations](./docs/limitations.md)         | Beta status and what this SDK does not do yet                                |
 
 ## Examples