> ## Documentation Index > Fetch the complete documentation index at: https://dev-docs.trychainshift.ai/llms.txt > Use this file to discover all available pages before exploring further. # Submit a run > 조직의 프롬프트를 지정한 모델로 실행합니다. 정기 모니터링이 아니라 요청한 시점에 한 번 수행됩니다. 요청이 성공하면 프롬프트 실행(Run)이 생성됩니다. 실행 대상은 promptIDs(개별 프롬프트) 또는 promptSetID(프롬프트 세트 전체) 중 하나만 지정할 수 있으며, 둘을 동시에 사용할 수 없습니다. 진행 상태는 응답으로 받은 id를 사용해 프롬프트 실행 상태 조회에서 확인합니다. ### 오류 | HTTP | 코드 | 설명 | | --- | --- | --- | | 400 | `COMMON_BINDING` | • 요청 본문 JSON 파싱 실패 • promptIDs·promptSetID 형식 오류 (details 에 `PLATFORM_INVALID_ID`) | | 400 | `COMMON_VALIDATION` | • models 누락·8개 초과·허용 목록 밖의 값 • promptIDs 500개 초과 • repeatCount 범위 위반(1~100) | | 400 | `RUN_PROMPT_TARGET_REQUIRED` | promptIDs·promptSetID 중 정확히 하나를 지정하지 않음 | | 400 | `PROMPT_SET_EMPTY` | 지정한 세트에 실행할 프롬프트가 없음 | | 400 | `MODEL_NOT_SUPPORTED` | 플랫폼 API 가 지원하지 않는 model 값 | | 400 | `MODEL_DISCONTINUED` | 서비스가 중단된 모델을 요청함 (현재 중단된 모델은 없음). 중단 모델은 `GET /models` 에도 나타나지 않음 | | 400 | `NO_AVAILABLE_PLATFORM` | 요청한 모델이 현재 모두 수집 불가 | | 400 | `REGION_NOT_SUPPORTED` | 지원하지 않는 지역 코드 | | 400 | `REGION_NOT_SUPPORTED_FOR_MODEL` | 모델이 지원하지 않는 지역 | | 400 | `LOCATION_NOT_SUPPORTED` | 지원하지 않는 위치 코드 (`GET /locations` 에 없는 값) | | 400 | `LOCATION_REGION_MISMATCH` | 지정한 region 에 속하지 않는 위치 | | 400 | `LOCATION_NOT_SUPPORTED_FOR_MODEL` | 요청한 모델 중 이 위치를 지원하지 않는 모델이 있음 | | 400 | `QUESTION_NOT_MARKETING_PLATFORM` | 플랫폼 API 소속이 아닌 프롬프트 | | 400 | `RUN_TOO_MANY_TASKS` | 총 실행 수(프롬프트 × 모델 × 반복)가 10,000 초과 | | 401 | `COMMON_UNAUTHORIZED` | API 키 또는 액세스 토큰이 없거나 유효하지 않음 | | 402 | `INSUFFICIENT_CREDITS` | 조직 크레딧 잔액이 이 요청의 단가보다 적음 | | 403 | `MODEL_NOT_GRANTED` | 별도 계약이 필요한 모델(`CLAUDE`, `PERPLEXITY`, `GROK`, `DEEPSEEK`)을 권한 없이 요청함 (`GET /models` 에도 나타나지 않음) | | 403 | `REGION_NOT_GRANTED` | 별도 계약이 필요한 지역을 권한 없이 요청함 (`GET /regions` 에도 나타나지 않음) | | 404 | `PROMPT_NOT_FOUND` | 프롬프트가 없거나 타 조직 소유 | | 404 | `PROMPT_SET_NOT_FOUND` | 세트가 없거나 타 조직 소유 | | 404 | `CAMPAIGN_NOT_FOUND` | 프롬프트가 속한 캠페인을 찾을 수 없음 | | 429 | `PLATFORM_RATE_LIMIT_EXCEEDED` | 요청 한도 초과. `Retry-After` 헤더의 초만큼 기다린 뒤 재시도 | | 500 | — | 서버 내부 오류. `code` 없이 `message` 와 `requestID` 를 반환 | See the [Running prompts](https://dev-docs.trychainshift.ai/en/guides/submitting-runs.md) guide for the full flow (prepare prompts → submit → poll → retrieve answers) and polling loop code. ## FAQ **Where do I get the results after submitting?** Poll [Get run status](https://dev-docs.trychainshift.ai/en/api-reference/runs/get-run-status.md) with the returned `runID`, and once it is `FINISHED`, query [List answers](https://dev-docs.trychainshift.ai/en/api-reference/answers/list-answers.md) with the `runID` filter. ## OpenAPI ````yaml openapi.en.json POST /v1/platform/runs openapi: 3.1.0 info: description: >- Platform API for external partners. Uses organization API key based authentication. ### Common request limits - The request body size is limited to 2MiB. Exceeding it returns 413 with the `COMMON_REQUEST_BODY_TOO_LARGE` code. - Array parameters and string fields have per-endpoint count and length limits. The limits are stated in each field or parameter description, and violating them returns 400 with the `COMMON_VALIDATION` code. ### Error responses The error response body always has the same three fields regardless of the status code. | Field | Type | Meaning | | --- | --- | --- | | `code` | string | A fixed identifier for the error type. This is the only value clients should branch on. The codes each endpoint returns are listed in the "Errors" table of that API. | | `message` | string | A human-readable, one-line summary of the whole error. There is always exactly one. | | `details` | array of strings | Individual reasons that do not fit in the single `message` line. Even a single reason is returned as an array of length 1, and the field is omitted when there is nothing to include. | - `details` is an array in every error. It is never returned as a string or an object, so no type-based branching is needed. - The wording of `details` is not a fixed contract. Use it only for display and debugging, and do not parse the strings to branch. Branch on `code` instead. - `details` is the only field that carries multiple reasons. There is no `messages` field in the response. - On validation failure (`COMMON_VALIDATION`), `details` contains one `"field: reason"` line per violated field. Example: `{"code":"COMMON_VALIDATION","message":"Data validation failed","details":["keywords: up to 100 items can be entered"]}` - `message` defaults to Korean and is returned in English when you send `Accept-Language: en`. `details` follows the same rule when it carries per-field reasons such as validation failures, but errors that include raw text for diagnostic purposes may contain wording in a language other than the requested one. - Only 500 has a different shape. It returns `message` and a `requestID` to reference when contacting support, without `code`. title: Platform API version: '1.0' servers: - url: http://dev-platform-api.trychainshift.ai/api security: [] tags: - description: >- The top-level account unit that API keys belong to. Retrieve the organization's basic information. name: Organization - description: >- A target brand whose exposure in AI services is monitored. Provides brand registration, retrieval, update, and deletion. name: Brand - description: >- Keyword intelligence for analyzing brands and markets. Provides Google and Naver search metrics, related and autocomplete keyword lookups, and site URL based keyword analysis. name: Keyword (Beta) - description: >- A group that organizes prompts by purpose. Provides create, get, update, and delete operations for prompt sets. name: PromptSet - description: >- A sentence-level unit of a query sent to an AI service. Provides prompt create, get, update, and delete, plus bulk creation. name: Prompt - description: AI service models used for prompt runs. Lists the supported models. name: Model - description: The region used to run prompts. Retrieves the list of supported regions. name: Region - description: >- The unit of work that actually runs prompts against AI services. Provides prompt run requests and retrieval of progress status and lists. name: Run - description: >- An AI service response collected by a run. Retrieves answers including sources and search queries. name: Answer - description: >- Statistics on the sources cited in answers, at the domain and URL level. Retrieve the citation summary and per-domain and per-URL citation statistics. name: SourceStatistics (Beta) - description: >- Statistics on tokens (Kiwi nouns) in answer bodies. Retrieves a ranking of the most frequent tokens and the tokens near a specific word. name: AnswerStatistics (Beta) - description: >- Statistics on the search queries (fan-outs) the AI actually issued while generating answers. Retrieves a ranking of search queries by occurrence count. name: FanoutStatistics (Beta) - description: >- An auxiliary feature that identifies proper nouns in text. Provides proper noun extraction. name: ProperNoun (Beta) - description: >- Organization credit balance and Platform API billing information. Retrieves the current credit balance. name: Credit externalDocs: description: '' url: '' paths: /v1/platform/runs: post: tags: - Run summary: Prompt run request description: >- 조직의 프롬프트를 지정한 모델로 실행합니다. 정기 모니터링이 아니라 요청한 시점에 한 번 수행됩니다. 요청이 성공하면 프롬프트 실행(Run)이 생성됩니다. 실행 대상은 promptIDs(개별 프롬프트) 또는 promptSetID(프롬프트 세트 전체) 중 하나만 지정할 수 있으며, 둘을 동시에 사용할 수 없습니다. 진행 상태는 응답으로 받은 id를 사용해 프롬프트 실행 상태 조회에서 확인합니다. ### 오류 | HTTP | 코드 | 설명 | | --- | --- | --- | | 400 | `COMMON_BINDING` | • 요청 본문 JSON 파싱 실패 • promptIDs·promptSetID 형식 오류 (details 에 `PLATFORM_INVALID_ID`) | | 400 | `COMMON_VALIDATION` | • models 누락·8개 초과·허용 목록 밖의 값 • promptIDs 500개 초과 • repeatCount 범위 위반(1~100) | | 400 | `RUN_PROMPT_TARGET_REQUIRED` | promptIDs·promptSetID 중 정확히 하나를 지정하지 않음 | | 400 | `PROMPT_SET_EMPTY` | 지정한 세트에 실행할 프롬프트가 없음 | | 400 | `MODEL_NOT_SUPPORTED` | 플랫폼 API 가 지원하지 않는 model 값 | | 400 | `MODEL_DISCONTINUED` | 서비스가 중단된 모델을 요청함 (현재 중단된 모델은 없음). 중단 모델은 `GET /models` 에도 나타나지 않음 | | 400 | `NO_AVAILABLE_PLATFORM` | 요청한 모델이 현재 모두 수집 불가 | | 400 | `REGION_NOT_SUPPORTED` | 지원하지 않는 지역 코드 | | 400 | `REGION_NOT_SUPPORTED_FOR_MODEL` | 모델이 지원하지 않는 지역 | | 400 | `LOCATION_NOT_SUPPORTED` | 지원하지 않는 위치 코드 (`GET /locations` 에 없는 값) | | 400 | `LOCATION_REGION_MISMATCH` | 지정한 region 에 속하지 않는 위치 | | 400 | `LOCATION_NOT_SUPPORTED_FOR_MODEL` | 요청한 모델 중 이 위치를 지원하지 않는 모델이 있음 | | 400 | `QUESTION_NOT_MARKETING_PLATFORM` | 플랫폼 API 소속이 아닌 프롬프트 | | 400 | `RUN_TOO_MANY_TASKS` | 총 실행 수(프롬프트 × 모델 × 반복)가 10,000 초과 | | 401 | `COMMON_UNAUTHORIZED` | API 키 또는 액세스 토큰이 없거나 유효하지 않음 | | 402 | `INSUFFICIENT_CREDITS` | 조직 크레딧 잔액이 이 요청의 단가보다 적음 | | 403 | `MODEL_NOT_GRANTED` | 별도 계약이 필요한 모델(`CLAUDE`, `PERPLEXITY`, `GROK`, `DEEPSEEK`)을 권한 없이 요청함 (`GET /models` 에도 나타나지 않음) | | 403 | `REGION_NOT_GRANTED` | 별도 계약이 필요한 지역을 권한 없이 요청함 (`GET /regions` 에도 나타나지 않음) | | 404 | `PROMPT_NOT_FOUND` | 프롬프트가 없거나 타 조직 소유 | | 404 | `PROMPT_SET_NOT_FOUND` | 세트가 없거나 타 조직 소유 | | 404 | `CAMPAIGN_NOT_FOUND` | 프롬프트가 속한 캠페인을 찾을 수 없음 | | 429 | `PLATFORM_RATE_LIMIT_EXCEEDED` | 요청 한도 초과. `Retry-After` 헤더의 초만큼 기다린 뒤 재시도 | | 500 | — | 서버 내부 오류. `code` 없이 `message` 와 `requestID` 를 반환 | operationId: submitPlatformRun requestBody: content: application/json: schema: $ref: '#/components/schemas/domain.SubmitRunRequest' description: Prompt run request required: true responses: '200': content: application/json: schema: $ref: '#/components/schemas/domain.SubmitRunResponse' description: OK '400': content: application/json: schema: $ref: '#/components/schemas/domain.ValidationError' description: >- models missing or more than 8, promptIDs more than 500, repeatCount out of range (1 to 100). For 400 codes related to run target, model, region, and location, see the error table above '401': content: application/json: schema: $ref: '#/components/schemas/domain.UnauthorizedError' description: API key or access token is missing or invalid '402': content: application/json: schema: $ref: '#/components/schemas/domain.InsufficientCreditsError' description: >- The credits required for the run size (models × prompts × repeats) exceed the balance '403': content: application/json: schema: $ref: '#/components/schemas/domain.ModelNotGrantedError' description: >- 별도 계약이 필요한 모델(MODEL_NOT_GRANTED) 또는 지역(REGION_NOT_GRANTED)입니다. 조직에 부여되지 않았습니다 '404': content: application/json: schema: $ref: '#/components/schemas/domain.PromptNotFoundError' description: >- The prompt (PROMPT_NOT_FOUND) or prompt set (PROMPT_SET_NOT_FOUND) was not found (including ones owned by another organization) '429': content: application/json: schema: $ref: '#/components/schemas/domain.RateLimitExceededError' description: >- Rate limit exceeded. Wait the number of seconds in the Retry-After header, then retry '500': content: application/json: schema: $ref: '#/components/schemas/domain.InternalServerError' description: >- Internal server error. Include the requestID in your report so it can be traced security: - OrgAPIKeyAuth: [] components: schemas: domain.SubmitRunRequest: properties: location: description: >- Run location (city-level, below the country). If omitted, the run uses the region (country) level. Only codes retrievable from GET /locations are allowed, and if region is also specified, the location must belong to that region. If region is omitted, the run uses the location's region. Every requested model must support this location (check with GET /locations?model=). example: US-NEW_YORK type: string models: description: >- List of models to run. Only values retrievable from GET /models are allowed. Discontinued models are rejected with 400 MODEL_DISCONTINUED before validation, so they are not kept in the oneof list (none at present) example: - CHATGPT - GEMINI items: $ref: '#/components/schemas/domain.RunModel' maxItems: 13 minItems: 1 type: array uniqueItems: false promptIDs: description: >- List of prompt IDs to run (1 to 500, duplicates count once). Specify exactly one of this or promptSetID example: - cM0RPoxbNbJy0mS-MhsHOg items: type: string maxItems: 500 minItems: 1 type: array uniqueItems: false promptSetID: description: >- ID of the prompt set to run: runs every prompt in the set. Specify exactly one of this and promptIDs. There is no 500-prompt limit on the set size; only the total run limit (10,000) applies example: IkVMHhjFWcBXcJ41-h9yBQ type: string region: description: >- Run region (ISO 3166-1 alpha-2). Defaults to KR when omitted. Unsupported codes return 400. Supported regions may differ by model (check with GET /regions?model=) example: KR type: string repeatCount: description: Number of repeated runs per prompt (question) (default 1, max 100). example: 1 maximum: 100 minimum: 1 type: integer required: - models type: object domain.SubmitRunResponse: properties: id: description: >- ID of the created prompt run. Pass this value as is as the runID of the run status endpoint (GET /v1/platform/runs/{runID}) example: S8cqDSjSot67bCK6YtmvHw type: string type: object domain.ValidationError: properties: code: description: Error code example: COMMON_VALIDATION type: string details: description: 'List of reasons per violating field (one "field: reason" per line)' example: - 'limit: must be 100 or less' items: type: string type: array uniqueItems: false message: description: >- Error message (Korean by default, English when Accept-Language: en is specified) example: Data validation failed type: string type: object domain.UnauthorizedError: properties: code: description: Error code example: COMMON_UNAUTHORIZED type: string message: description: >- Error message (Korean by default, English when Accept-Language: en is specified) example: Unauthenticated user type: string type: object domain.InsufficientCreditsError: properties: code: description: Error code example: INSUFFICIENT_CREDITS type: string message: description: >- Error message (Korean by default, English when Accept-Language: en is specified) example: Insufficient credits. type: string type: object domain.ModelNotGrantedError: properties: code: description: Error code example: MODEL_NOT_GRANTED type: string message: description: >- Error message (Korean by default, English when Accept-Language: en is specified) example: >- This model is not granted to the organization. An agreement is required to use it type: string type: object domain.PromptNotFoundError: properties: code: description: Error code example: PROMPT_NOT_FOUND type: string message: description: >- Error message (Korean by default, English when Accept-Language: en is specified) example: Prompt not found type: string type: object domain.RateLimitExceededError: properties: code: description: Error code example: PLATFORM_RATE_LIMIT_EXCEEDED type: string message: description: >- Error message (Korean by default, English when Accept-Language: en is specified) example: Too many requests. Please try again later. type: string type: object domain.InternalServerError: properties: message: description: Error message (not localized) example: Internal Server Error type: string requestID: description: >- Request identifier to reference in support inquiries (same value as the X-Request-Id response header) example: bkZqnHQepkgTOocTvyHWSlHcGzRqQbZk type: string type: object domain.RunModel: enum: - CHATGPT - GOOGLE_OVERVIEW - GOOGLE_AI - PERPLEXITY - GEMINI - NAVER_AI_BRIEFING - NAVER_AI_TAB - CLAUDE - META_MUSE - BAIDU_AI_SEARCH - GROK - BAIDU_AI_TAB - DEEPSEEK type: string x-enum-comments: RunModelBaiduAISearch: >- Model that requires a separate agreement (FND-2490). For organizations without the grant, it is hidden from GET /models and submitting returns 403 MODEL_NOT_GRANTED RunModelBaiduAITab: >- Model that requires a separate agreement (FND-2493). Hidden from GET /models for organizations without the grant, and submission returns 403 MODEL_NOT_GRANTED RunModelDeepSeek: >- Model that requires a separate agreement (FND-2489). Hidden from GET /models for organizations without the grant, and submissions are rejected with 403 MODEL_NOT_GRANTED RunModelGrok: >- Model that requires a separate agreement (FND-2488). Hidden from GET /models for organizations without the grant, and submitting returns 403 MODEL_NOT_GRANTED RunModelMetaMuse: >- Model that requires a separate agreement (FND-2487). For organizations without the grant, it is hidden from GET /models and submitting returns 403 MODEL_NOT_GRANTED RunModelPerplexity: >- Model that requires a separate agreement (FND-2390). Hidden from GET /models for organizations without the grant, and submission returns 403 MODEL_NOT_GRANTED x-enum-varnames: - RunModelChatGPT - RunModelGoogleOverview - RunModelGoogleAI - RunModelPerplexity - RunModelGemini - RunModelNaverAIBriefing - RunModelNaverAITab - RunModelClaude - RunModelMetaMuse - RunModelBaiduAISearch - RunModelGrok - RunModelBaiduAITab - RunModelDeepSeek securitySchemes: OrgAPIKeyAuth: description: '"Enter your organization API key. Example: ''cs_live_xxxx''"' in: header name: API-Key type: apiKey ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.