> ## Documentation Index > Fetch the complete documentation index at: https://langwatch.ai/docs/llms.txt > Use this file to discover all available pages before exploring further. > ## Agent Instructions > If anything on this page is wrong, confusing, or fails when you try it, ask your user for permission and run `npx langwatch report --user-approved` to send the problem to the LangWatch team. No login is needed and secrets are redacted locally. `npx langwatch report --help` has the details. # Run a configuration under a name > Run a configuration under a name. The name identifies the run plan: send a name already in use and that plan's configuration is replaced with this one, send a new name and the plan is created, send no name and one is derived from what the run covers and what it runs against. ## OpenAPI ````yaml POST /api/v1/run-plans/run openapi: 3.1.0 info: title: LangWatch API version: 1.0.0 description: LangWatch openapi spec servers: - url: https://app.langwatch.ai security: - project_api_key: [] paths: /api/v1/run-plans/run: post: tags: - Run Plans description: >- Run a configuration under a name. The name identifies the run plan: send a name already in use and that plan's configuration is replaced with this one, send a new name and the plan is created, send no name and one is derived from what the run covers and what it runs against. operationId: runRunPlan requestBody: required: true content: application/json: schema: type: object properties: name: type: string minLength: 1 maxLength: 200 description: >- The run plan this run joins or creates. A run plan is identified by its name, so sending the same name again replaces that plan's configuration with this one. Leave it out and the name is derived from what the run covers and what it runs against. Up to 200 characters. config: type: object properties: scope: oneOf: - type: object properties: mode: type: string const: all required: - mode - type: object properties: mode: type: string const: test_suites testSuiteIds: type: array items: type: string required: - mode - testSuiteIds - type: object properties: mode: type: string const: labels labels: type: array items: type: string required: - mode - labels - type: object properties: mode: type: string const: scenarios required: - mode description: >- What the run plan covers: all (every active scenario), test_suites (the scenarios filed in the named test suites), labels (the scenarios carrying any of the labels), or scenarios (the scenarioIds sent with the configuration). A dynamic scope is resolved again at every run, so a scenario written later runs without editing the plan. targets: type: array items: type: object properties: type: type: string enum: - prompt - http - code - workflow - connected description: >- What kind of thing the scenarios run against. A connected agent is one registered from code with the SDK. referenceId: type: string description: >- The id of the prompt, agent or workflow to run against. A connected target may also say @, for example support-agent@production, which resolves to the agent id. runParameters: type: object additionalProperties: anyOf: - type: string maxLength: 4096 - type: number - type: boolean description: >- Parameter values this target alone runs with, by name. They are merged over the run-level parameters and the target wins, so two targets may name the same agent with different values: that is how one run compares one agent on two models, and the results show one column for each target. required: - type - referenceId description: >- The prompts, agents or workflows every scenario runs against. Every target runs every scenario, so naming more than one compares them in the same run. repeatCount: type: integer minimum: 1 maximum: 5 description: >- How many times each scenario and target pairing runs. Between 1 and 5; defaults to 1. simulatorModel: type: - string - 'null' pattern: ^[a-zA-Z0-9_-]+(?:\/[^\s/]+)+$ maxLength: 200 description: >- The model that plays the user for every scenario in the run. Overrides each scenario's own choice. Leave it out for the scenario or project default. judgeModel: type: - string - 'null' pattern: ^[a-zA-Z0-9_-]+(?:\/[^\s/]+)+$ maxLength: 200 description: >- The model that judges every scenario in the run. Overrides each scenario's own choice. Leave it out for the scenario or project default. scenarioIds: type: array items: type: string description: >- The scenarios a test_suites or scenarios scope covers. Read by a scenarios scope alone; a scope that states a rule resolves its own list at run time. required: - scope - targets description: >- What this run covers and what it runs against. Written onto the run plan the name resolves. idempotencyKey: type: string description: >- Repeat the same key to make a retry join the batch the first call started instead of running everything again. Defaults to a new key per call. parameters: type: object additionalProperties: anyOf: - type: string maxLength: 4096 - type: number - type: boolean description: >- Constant values applied to every scenario in the run, e.g. a fixture id or a tenant. A value supplied here overrides the scenario's own default for that name, and a target that names the same parameter in its runParameters overrides it for that target. note: type: string maxLength: 200 description: >- One short line describing why this batch was run, e.g. a commit hash or what you changed. It is stored on every run of the batch and shown beside the run in the platform. Up to 200 characters. required: - config responses: '200': description: Success content: application/json: schema: type: object properties: scheduled: type: boolean description: True once the runs are queued. batchRunId: type: string description: The id of this batch. Every run started here carries it. setId: type: string description: The result set the batch is filed under in the platform. jobCount: type: number description: How many runs were queued. skippedArchived: type: object properties: scenarios: type: array items: type: string description: Scenarios left out because they are archived. targets: type: array items: type: string description: Targets left out because they are archived. required: - scenarios - targets description: What the run left out, and why. items: type: array items: type: object properties: scenarioRunId: type: string description: The id of this single run. scenarioId: type: string description: The scenario that was run. target: type: object properties: type: type: string enum: - prompt - http - code - workflow - connected description: >- What kind of thing the scenarios run against. A connected agent is one registered from code with the SDK. referenceId: type: string description: >- The id of the prompt, agent or workflow to run against. A connected target may also say @, for example support-agent@production, which resolves to the agent id. runParameters: type: object additionalProperties: anyOf: - type: string maxLength: 4096 - type: number - type: boolean description: >- Parameter values this target alone runs with, by name. They are merged over the run-level parameters and the target wins, so two targets may name the same agent with different values: that is how one run compares one agent on two models, and the results show one column for each target. required: - type - referenceId description: What it was run against. name: type: - string - 'null' description: The scenario name, when known. required: - scenarioRunId - scenarioId - target - name description: Every run this call queued. runPlanId: type: string description: The run plan this run was filed under. planName: type: string description: The name that plan answers to. created: type: boolean description: >- True when this run created the plan, false when it joined a plan already there. platformUrl: type: string format: uri description: Where to watch this run in the LangWatch platform. required: - scheduled - batchRunId - setId - jobCount - skippedArchived - items - runPlanId - planName - created - platformUrl security: - project_api_key: [] components: securitySchemes: project_api_key: type: apiKey in: header name: X-Auth-Token description: >- Project API key for sending traces and accessing project-scoped resources. Format: sk-lw-... (no underscore). Obtain one by creating a project via the Admin API or the LangWatch UI. ````