> ## Documentation Index > Fetch the complete documentation index at: https://docs.spidra.io/llms.txt > Use this file to discover all available pages before exploring further. # Retry a template run > Runs the failed pages of a finished run again, with the run's original inputs, and leaves the rest alone. The run goes back to `running`, so poll [Get a template run](/api-reference/templates/get-run) as before. Each failure in `failures` has a `code` (such as `blocked` or `not_found`) to decide whether a retry is worth it. Works on runs of a single step. Runs with a second step (such as checking each website for contacts) or that run many searches answer `409` with `RETRY_NOT_SUPPORTED`; start a new run instead. **Credits:** only the retried pages, at the template's usual price. A page that fails again costs nothing. ## OpenAPI ````yaml POST /template-runs/{runId}/retry openapi: 3.1.0 info: title: Spidra API version: 1.0.0 description: >- Public API endpoints for web scraping via Spidra. Authenticate with `Authorization: Bearer YOUR_API_KEY` or `x-api-key: YOUR_API_KEY`; both work on every endpoint. servers: - url: https://api.spidra.io/api security: - BearerAuth: [] - ApiKeyAuth: [] paths: /template-runs/{runId}/retry: post: tags: - Templates summary: Retry a template run's failed pages description: > Runs the failed pages of a finished run again, with the run's original inputs, and leaves the rest alone. The run goes back to `running`, so poll [Get a template run](/api-reference/templates/get-run) as before. Each failure in `failures` has a `code` (such as `blocked` or `not_found`) to decide whether a retry is worth it. Works on runs of a single step. Runs with a second step (such as checking each website for contacts) or that run many searches answer `409` with `RETRY_NOT_SUPPORTED`; start a new run instead. **Credits:** only the retried pages, at the template's usual price. A page that fails again costs nothing. parameters: - name: runId in: path required: true description: The `runId` returned when the run started. schema: type: string format: uuid example: dfefa844-ad1f-4400-b127-dc17a186d9c2 requestBody: required: false content: application/json: schema: type: object properties: cookies: type: string description: >- Cookies are never stored. Send them again if the run used them. responses: '202': description: The failed pages are running again. content: application/json: schema: type: object properties: status: type: string enum: - success retriedCount: type: integer description: How many pages are running again. warnings: type: array items: type: string run: $ref: '#/components/schemas/TemplateRun' example: status: success retriedCount: 1 run: runId: cba40eaa-3593-41c2-8688-d0984d255695 template: github-repo-details status: running '400': description: '`NO_FAILED_PAGES`: the run has no failed pages.' '404': description: No run with this ID. '409': description: >- `RUN_RUNNING`: the run hasn't finished yet. `RETRY_NOT_SUPPORTED`: the run has more than one step. components: schemas: TemplateRun: type: object description: A template run, with its rows once it has finished. properties: runId: type: string format: uuid description: The run's ID. Use it with `GET /template-runs/{runId}`. template: type: string description: The template's slug. templateVersion: type: integer description: The template version the run used. presetId: type: - string - 'null' description: The preset the run was started from, or `null`. tool: type: string enum: - batch - search - map - crawl - parse description: The Spidra tool behind the run. status: type: string enum: - running - completed - partial - failed - cancelled description: > - `running`: still reading pages. - `completed`: finished. `rows` holds the results. - `partial`: some pages couldn't be read, or a second step failed. The pages that failed are in `failures` and aren't charged. `rows` holds the rest. - `failed`: nothing could be read. `error` says why, and you aren't charged. - `cancelled`: the job behind the run was cancelled. inputs: type: object description: The inputs the run used, with defaults filled in. createdAt: type: string format: date-time description: When the run started. finishedAt: type: - string - 'null' format: date-time description: When the run finished, or `null` while it runs. creditsUsed: type: - number - 'null' description: Credits the run cost. `null` until it has finished. urlCount: type: - integer - 'null' description: How many pages or searches the run started with. rowCount: type: - integer - 'null' description: How many rows the run returned. progress: type: object description: >- Pages done so far. With a second step, its pages are added once they're queued. properties: completed: type: integer description: Pages read. failed: type: integer description: Pages that couldn't be read. total: type: integer description: Pages known so far. rows: type: array items: type: object description: >- The results, one object per row, once the run has finished. Each row has every field the site returned, not only the template's default `columns`. Left out with `?rows=false`. failures: type: array description: Pages that couldn't be read. They aren't charged. items: type: object properties: url: type: string description: The page. error: type: string description: Why it couldn't be read. code: type: string enum: - not_found - no_results - login_required - blocked - timeout - empty_content - invalid_input - unsupported - out_of_credits - failed description: > Why the page failed, as a stable code to branch on: - `not_found`: the link points at nothing (no such page, company or place). - `no_results`: the search or list exists but has nothing matching. - `login_required`: only signed-in users can see it. Spidra reads public pages only. - `blocked`: the site refused or challenged every attempt. - `timeout`: the page took too long. - `empty_content`: the page loaded but held no usable data. - `invalid_input`: the input can't be used as given. - `unsupported`: this kind of page isn't read. - `out_of_credits` - `failed`: anything else. error: type: string description: >- Only when the run failed, or ended `partial` because its second step failed. Why. followStep: type: object description: >- Only on templates with a second step switched on, such as **Find emails and social links**. The run stays `running` until this step has finished, and its fields are merged into each row. properties: status: type: string enum: - waiting - running - completed - failed description: Where the second step is. sites: type: integer description: Websites the step checks. pagesLoaded: type: integer description: Pages it read. pagesFailed: type: integer description: Pages it couldn't read. These aren't charged. pagesWaiting: type: integer description: Pages still to read. creditsUsed: type: integer description: Credits the step cost. error: type: string description: Only when the step failed. Why. summary: type: object description: >- Only on rank trackers. How many keywords or points ranked, and the average position. job: type: object description: >- The tool job behind the run. For crawl and parse templates, the full result is at that tool's own endpoint, such as [Get a crawl](/api-reference/crawling/crawl-status). properties: tool: type: string enum: - batch - search - map - crawl - parse description: The tool. id: type: - string - 'null' description: The tool job's ID, or `null` when there's none to read. resultsPath: type: string description: >- Only for crawl and parse templates. The dashboard page with the full result. securitySchemes: BearerAuth: type: http scheme: bearer bearerFormat: API key ApiKeyAuth: type: apiKey in: header name: x-api-key ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.