Bring Structured Content
to OpenAI Agents SDK

Provide the dataset name (e.g. "staging", "development"). The dataset will be empty and ready for content. Dataset names must be lowercase alphanumeric. Create a new Sanity dataset

Requires a JSON object with _type and content fields. For example: {"_type": "post", "title": "Hello", "body": "World"}. Returns the created document with its generated ID. Create a new Sanity document

Provide the document ID. WARNING: this action is irreversible. Delete a Sanity document

g. "drafts.abc123" or "abc123"). Returns the full document with all fields. Useful for inspecting individual content items. Get a specific Sanity document by ID

Datasets are isolated content collections within a project (e.g. "production", "staging", "development"). List all datasets in the Sanity project

g. "post", "product", "author") from the configured dataset. Each document returns its _id, content fields and metadata. Use this to browse content collections. List documents of a specific type in Sanity

Each asset includes its _id, original filename, dimensions and metadata. Optionally set a limit. Useful for managing media library content. List image assets in Sanity

Useful for auditing project access and team management. List project members in Sanity

Requires a GROQ query string. Optionally provide params as JSON for parameterized queries. Returns query results as an array. Example query: "*[_type == 'post' && publishedAt > $date]{title, slug}" Run a GROQ query against Sanity

Optionally filter by document types and set a result limit. Returns matching documents with their full content. Search documents in Sanity

Requires the document ID and a JSON object with the fields to set. For example: {"title": "Updated Title", "status": "published"}. Only provided fields will be modified. Update fields on a Sanity document