> ## Documentation Index > Fetch the complete documentation index at: https://docs.lehar.ai/llms.txt > Use this file to discover all available pages before exploring further. # Create test suite > Creates a reusable suite of persona scenarios for the Test Runner. `agent_id` is optional — a suite is a workspace-level object and the agent to test is chosen at run time; when provided it is validated against the workspace. A scenario's persona may be left blank while drafting and is required only at run time. Scope `sessions:write`. ## OpenAPI ````yaml /api-reference/openapi.json post /agent-test-suites openapi: 3.1.0 info: title: Lehar API version: v0 description: >- Customer-facing REST API for the Lehar voice-AI platform. Platform-admin endpoints are intentionally excluded from this reference. Regenerate with `npm run sync:openapi`. servers: - url: https://api.lehar.ai/ca/api/v0 security: - ApiKeyAuth: [] - BearerAuth: [] paths: /agent-test-suites: post: tags: - Agent Testing summary: Create test suite description: >- Creates a reusable suite of persona scenarios for the Test Runner. `agent_id` is optional — a suite is a workspace-level object and the agent to test is chosen at run time; when provided it is validated against the workspace. A scenario's persona may be left blank while drafting and is required only at run time. Scope `sessions:write`. operationId: create_suite_ca_api_v0_agent_test_suites_post requestBody: required: true content: application/json: schema: type: object required: - name properties: name: type: string maxLength: 200 description: Suite name, unique within the workspace (max 200 chars). agent_id: type: string description: >- Optional agent to associate as an origin hint. Validated against the workspace when set. scenarios: type: array maxItems: 100 description: >- Persona scenarios (max 100). Most fields are passed through to the Test Runner; only `runs`, `max_turns`, and `variables` are validated here. items: type: object properties: label: type: string description: >- Display label; becomes the result's `scenario_label` (defaults to `scenario N`). instructions: type: string description: >- Persona / instructions that drive the simulated user (alias: `persona`). A non-empty `instructions` or `persona` is required to run the suite. persona: type: string description: Alias for `instructions`. agent_expectations: type: string description: What the agent is expected to do, for the judges. judges: type: array items: type: string description: >- Judge criteria for this scenario (overrides the suite `config.judges`). tool_mocks: type: object description: >- Mocked tool responses, passed opaquely to the Test Runner. runs: type: integer minimum: 1 maximum: 5 default: 1 description: How many times to repeat this scenario (1-5). max_turns: type: integer minimum: 1 maximum: 20 default: 8 description: Max conversation turns per run (1-20). simulator_model: type: string description: >- Model driving the simulated user (e.g. `gemini:gemini-2.5-flash`). judge_model: type: string description: Model used to judge the conversation. variables: type: object description: >- Per-scenario `{{variable}}` values. Flat map of scalar values (string/number/boolean/null); at most 64 keys and 30000 serialized bytes. Overrides suite `config.variables`. config: type: object description: >- Suite-level defaults. `variables` (validated) seeds `{{variable}}` values for every scenario; `judges`, `runs`, `max_turns`, `simulator_model`, `judge_model` act as defaults a scenario can override; `max_parallel` (1-8, default 3) bounds scenario fan-out. properties: variables: type: object description: >- Default `{{variable}}` values. Same shape and caps as a scenario's `variables`. judges: type: array items: type: string max_parallel: type: integer minimum: 1 maximum: 8 default: 3 example: name: Renewal objections agent_id: agent_sales_hi scenarios: - label: Price objection instructions: >- You are a budget-conscious customer who thinks the renewal price is too high. Push back on cost before agreeing to anything. runs: 2 max_turns: 8 config: judges: - task_completion - safety variables: company: Acme responses: '201': description: Created test suite content: application/json: example: id: test_suite_01hzy8x2m4q7r8s9t0v1w2x3y4 agent_id: agent_sales_hi name: Renewal objections scenarios: - label: Price objection instructions: >- You are a budget-conscious customer who thinks the renewal price is too high. Push back on cost before agreeing to anything. runs: 2 max_turns: 8 config: judges: - task_completion - safety variables: company: Acme created_by_user_id: user_01EXAMPLE created_at: '2026-05-22T10:00:00Z' updated_at: '2026-05-22T10:00:00Z' '400': description: >- Validation failed — e.g. missing/oversized `name`, `scenarios` not a list or over 100, a scenario `runs`/`max_turns` out of range, invalid `variables`, `config` not an object, or `agent_id` not in the workspace. content: application/json: example: error: code: invalid_request message: name is required '401': description: Missing or invalid credentials (code `unauthorized`). '403': description: >- The credential is missing the required `sessions:write` scope (code `forbidden`). '409': description: A suite with this name already exists in the workspace. content: application/json: example: error: code: conflict message: >- A test suite named 'Renewal objections' already exists in this workspace '500': description: Unexpected server error (code `internal_error`). components: securitySchemes: ApiKeyAuth: type: apiKey in: header name: X-API-KEY BearerAuth: type: http scheme: bearer ````