> ## Documentation Index > Fetch the complete documentation index at: https://docs.base44.com/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > Internal links on these pages omit the .md extension. Append .md to a docs page URL, or send an Accept: text/markdown header, to get that page as markdown. # Seed test data > This API is in beta. Endpoints, fields, and behavior may still change, so avoid depending on it in production. Fills the app's test data with sample records for the testing agent to work with. A run on the app's preview uses test data, and [Run test](/api-reference/run-test) seeds it when such a run starts, so call this only to refresh the test data yourself. For each entity other than `User`: - When the entity has live records, up to 5 copies of them are written to test data, including any personal details they hold. Live data is only read, never changed. - Otherwise, for up to 5 such entities, AI makes up to 3 records each from the entity's schema. This is skipped for an entity whose test data has records someone added by hand, and for an entity with no schema fields. New records replace the ones an earlier seed wrote. Test records someone added by hand are kept. Copying is free. Making up records costs credits, and the call is refused when it needs to make some up and the workspace is out of credits, though records already copied are kept. It runs while you wait, usually for 20 to 45 seconds and up to a few minutes on an app with many entities, so allow a long timeout. Seeds of one app run one at a time. A call made while another is running waits for it and then usually seeds again, replacing what the first one wrote and charging again for any AI-made records. If the wait runs too long, it gives up and reports no records written. This is limited to 30 requests per minute per caller for each app. Some workspaces have a different limit. This endpoint accepts a personal API key belonging to a user with editor access to the app. Read-only keys and workspace API keys are refused. The response includes fields beyond the ones documented here. Don't rely on undocumented response fields, as they can change at any time. ## OpenAPI ````yaml /developers/references/app-management/app-management-openapi.json post /api/apps/{app_id}/testing-agent/data/seed openapi: 3.1.0 info: title: Base44 App Management API version: 1.0.0 servers: - url: https://app.base44.com security: - PersonalAccessTokenAuth: [] paths: /api/apps/{app_id}/testing-agent/data/seed: post: summary: Seed test data description: >- This API is in beta. Endpoints, fields, and behavior may still change, so avoid depending on it in production. Fills the app's test data with sample records for the testing agent to work with. A run on the app's preview uses test data, and [Run test](/api-reference/run-test) seeds it when such a run starts, so call this only to refresh the test data yourself. For each entity other than `User`: - When the entity has live records, up to 5 copies of them are written to test data, including any personal details they hold. Live data is only read, never changed. - Otherwise, for up to 5 such entities, AI makes up to 3 records each from the entity's schema. This is skipped for an entity whose test data has records someone added by hand, and for an entity with no schema fields. New records replace the ones an earlier seed wrote. Test records someone added by hand are kept. Copying is free. Making up records costs credits, and the call is refused when it needs to make some up and the workspace is out of credits, though records already copied are kept. It runs while you wait, usually for 20 to 45 seconds and up to a few minutes on an app with many entities, so allow a long timeout. Seeds of one app run one at a time. A call made while another is running waits for it and then usually seeds again, replacing what the first one wrote and charging again for any AI-made records. If the wait runs too long, it gives up and reports no records written. This is limited to 30 requests per minute per caller for each app. Some workspaces have a different limit. This endpoint accepts a personal API key belonging to a user with editor access to the app. Read-only keys and workspace API keys are refused. The response includes fields beyond the ones documented here. Don't rely on undocumented response fields, as they can change at any time. operationId: seed_data_api_apps__app_id__testing_agent_data_seed_post parameters: - name: app_id in: path required: true schema: type: string description: ID of the app. title: App Id description: ID of the app. example: 6820f3a4e7b91d003c45a1f2 responses: '200': description: What was written into the app's test data. content: application/json: schema: $ref: '#/components/schemas/SeededTestData' '400': description: >- Some entities need AI-made records and the workspace is out of credits. '401': description: Missing or invalid credentials. '403': description: >- You don't have editor access to this app, the app is blocked, or your API key is read-only or a workspace API key. '404': description: App not found. '409': description: Your workspace requires an unlocked SSO session. '429': description: Rate limit exceeded. components: schemas: SeededTestData: properties: seeded: additionalProperties: type: integer type: object title: Seeded description: >- Records written for each entity, keyed by entity name. `0` when nothing was written. example: Project: 5 Task: 3 total: type: integer title: Total description: Records written across all entities. example: 8 generated_total: type: integer title: Generated Total description: >- How many of the records were made up by AI rather than copied from live data. example: 3 entities_with_data: type: integer title: Entities With Data description: Entities that got records, or whose own test records were kept. example: 2 entities_without_prod_data: items: type: string type: array title: Entities Without Prod Data description: >- Entities that got nothing copied from live data and didn't report an error. When this call gave up waiting for another seed of the app, every entity is listed. example: - Task entities_with_errors: items: type: string type: array title: Entities With Errors description: >- Entities whose seeding reported an error, for example because generating records timed out. An entity whose record writes all failed isn't listed. Its `seeded` count is `0`. example: [] type: object required: - seeded - total - generated_total - entities_with_data - entities_without_prod_data - entities_with_errors title: SeededTestData description: What a seed wrote into the app's test data. securitySchemes: PersonalAccessTokenAuth: type: http scheme: bearer description: 'Personal access token, sent as `Authorization: Bearer `.' ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.