> ## Documentation Index > Fetch the complete documentation index at: https://docs.devtune.ai/llms.txt > Use this file to discover all available pages before exploring further. # Get Agent Picks run trace > Read one page of a project-owned run's console trace: redacted prompt, setup, narration, tool calls and results, file edits and errors, ordered by attempt then seq. An attempt's lines are available once that attempt finishes, so a running attempt has none yet; an attempt whose transcript could not be read has one error line, Trace unavailable for this attempt. availability says whether the run has stored lines, has none yet (pending: it is still running or finished within the last hour), or has none and none is expected (not_retained: it finished over an hour ago, as every run before traces were kept did), and attempts marks each attempt stored or not. A model_call item follows the first line of its attempt carrying its provider response id, else one of its tool-call ids, with tokens, cost and the running total of placed calls. Calls that match no line are in unplaced, 200 at a time in finish order: the first page (cursor 0, 0) carries the first 200 with unplaced_total, and passing unplaced_pagination.after_id as after_unplaced returns the next 200; other pages carry none. Each has a reason: failed, no_ids, pending while the run is running, or unmatched. Codex reports no ids, so its calls are never placed. totals counts every model proxy call. Each line carries preview, its first 65,536 characters, with preview_truncated when the stored text is longer; full=true with limit=1 returns one line's whole stored text (up to 1 MiB, detail.truncated beyond). Pass pagination.after_attempt and after_seq back until has_more is false. limit defaults to 50, maximum 100; a page also stops once 256 KiB of lines are included and always holds at least one line. Requires Agent Picks and visibility.read. ## OpenAPI ````yaml /openapi.json get /projects/{projectId}/agent-picks/runs/{runId}/trace openapi: 3.1.0 info: title: DevTune API version: 2.0.0 description: >- API for programmatic access to your AI visibility data, webhook subscriptions, and automation workflows. Use this API to integrate DevTune data into CI/CD pipelines, BI tools, AI agents, and operational systems. servers: - url: https://devtune.ai/api/v2 description: Production security: - bearerAuth: [] paths: /projects/{projectId}/agent-picks/runs/{runId}/trace: get: summary: Get Agent Picks run trace description: >- Read one page of a project-owned run's console trace: redacted prompt, setup, narration, tool calls and results, file edits and errors, ordered by attempt then seq. An attempt's lines are available once that attempt finishes, so a running attempt has none yet; an attempt whose transcript could not be read has one error line, Trace unavailable for this attempt. availability says whether the run has stored lines, has none yet (pending: it is still running or finished within the last hour), or has none and none is expected (not_retained: it finished over an hour ago, as every run before traces were kept did), and attempts marks each attempt stored or not. A model_call item follows the first line of its attempt carrying its provider response id, else one of its tool-call ids, with tokens, cost and the running total of placed calls. Calls that match no line are in unplaced, 200 at a time in finish order: the first page (cursor 0, 0) carries the first 200 with unplaced_total, and passing unplaced_pagination.after_id as after_unplaced returns the next 200; other pages carry none. Each has a reason: failed, no_ids, pending while the run is running, or unmatched. Codex reports no ids, so its calls are never placed. totals counts every model proxy call. Each line carries preview, its first 65,536 characters, with preview_truncated when the stored text is longer; full=true with limit=1 returns one line's whole stored text (up to 1 MiB, detail.truncated beyond). Pass pagination.after_attempt and after_seq back until has_more is false. limit defaults to 50, maximum 100; a page also stops once 256 KiB of lines are included and always holds at least one line. Requires Agent Picks and visibility.read. operationId: getAgentPicksRunTrace parameters: - name: projectId in: path required: true schema: type: string format: uuid - name: runId in: path required: true schema: type: string format: uuid - name: after_attempt in: query schema: type: integer minimum: 0 maximum: 2147483647 default: 0 - name: after_seq in: query schema: type: integer minimum: 0 maximum: 2147483647 default: 0 - name: limit in: query schema: type: integer minimum: 1 maximum: 100 default: 50 - name: full in: query description: >- Adds each line's whole stored text as text. Only with limit=1: pass the line's attempt and its seq minus one as the cursor. schema: type: boolean default: false - name: after_unplaced in: query description: >- The unplaced_pagination.after_id of the previous page; returns the next 200 unplaced calls. schema: type: string format: uuid responses: '200': description: Agent Picks run trace content: application/json: schema: type: object properties: data: $ref: '#/components/schemas/AgentPicksRunTrace' meta: $ref: '#/components/schemas/Meta' required: - data - meta '400': $ref: '#/components/responses/BadRequestError' '401': $ref: '#/components/responses/UnauthorizedError' '403': $ref: '#/components/responses/ForbiddenError' '404': $ref: '#/components/responses/NotFoundError' '429': $ref: '#/components/responses/RateLimitExceededError' components: schemas: AgentPicksRunTrace: type: object additionalProperties: false required: - availability - attempts - items - unplaced - unplaced_total - unplaced_pagination - totals - pagination properties: availability: type: string enum: - stored - pending - not_retained description: >- stored: the run has stored lines. pending: none yet, because the run is still running or finished within the last hour. not_retained: none, and none is expected, because the run finished over an hour ago, as every run before traces were kept did. attempts: type: array description: >- Each attempt the run started, and whether its lines are stored. An attempt with stored false has no trace. items: type: object additionalProperties: false required: - attempt - stored properties: attempt: type: integer minimum: 1 stored: type: boolean items: type: array items: type: object additionalProperties: false required: - attempt - seq - at - kind - preview - preview_truncated - detail properties: attempt: type: integer minimum: 1 seq: type: integer minimum: 1 description: >- The line within its attempt. A model_call item carries the seq of the line it follows. at: type: string format: date-time kind: type: string enum: - prompt - narration - tool_call - tool_result - file_write - file_edit - error - model_call preview: type: string description: >- The line's first 65,536 characters, redacted. Empty for a model_call item. preview_truncated: type: boolean description: The stored text is longer than the preview. text: type: string description: >- The line's whole stored text, up to 1 MiB (detail.truncated beyond). Only with full=true. detail: type: object additionalProperties: true description: >- Tool name, arguments summary, path, command, URL, status, exit code, line counts, packages and provider ids when the agent reported them. A model_call item has model_call_id, response_id, tool_call_ids, status, tokens, cost_usd, started_at, finished_at, running_total_usd and spend_is_lower_bound. unplaced: type: array maxItems: 500 items: type: object additionalProperties: false required: - model_call_id - attempt - reason - response_id - tool_call_ids - status - input_tokens - output_tokens - cache_read_tokens - cache_write_tokens - cost_usd - started_at - finished_at properties: model_call_id: type: string format: uuid attempt: type: - integer - 'null' reason: type: string enum: - failed - no_ids - pending - unmatched response_id: type: - string - 'null' tool_call_ids: type: array items: type: string status: type: string enum: - completed - failed input_tokens: type: integer output_tokens: type: integer cache_read_tokens: type: integer cache_write_tokens: type: integer cost_usd: type: - number - 'null' started_at: type: string format: date-time finished_at: type: string format: date-time unplaced_total: type: - integer - 'null' minimum: 0 description: Null on a page that carries no unplaced calls. unplaced_pagination: type: - object - 'null' additionalProperties: false required: - after_id - has_more description: >- Null on a page that carries no unplaced calls. Pass after_id as after_unplaced for the next 200. properties: after_id: type: - string - 'null' format: uuid has_more: type: boolean totals: type: object additionalProperties: false description: Every model proxy call of the run, placed or not. required: - calls_started - calls_finished - input_tokens - output_tokens - cache_read_tokens - cache_write_tokens - cost_usd - unpriced_calls properties: calls_started: type: integer calls_finished: type: integer input_tokens: type: integer output_tokens: type: integer cache_read_tokens: type: integer cache_write_tokens: type: integer cost_usd: type: number unpriced_calls: type: integer pagination: type: object additionalProperties: false required: - after_attempt - after_seq - has_more - limit properties: after_attempt: type: integer after_seq: type: integer has_more: type: boolean limit: type: integer Meta: type: object properties: timestamp: type: string format: date-time projectId: type: string format: uuid Error: type: object properties: error: type: string message: type: string reason: type: string description: >- Why the request was refused, for an error with more than one cause, such as a refused content-draft start. status: type: integer required: - error - message - status responses: BadRequestError: description: Bad request content: application/json: schema: $ref: '#/components/schemas/Error' UnauthorizedError: description: Unauthorized content: application/json: schema: $ref: '#/components/schemas/Error' ForbiddenError: description: Forbidden content: application/json: schema: $ref: '#/components/schemas/Error' NotFoundError: description: Not found content: application/json: schema: $ref: '#/components/schemas/Error' RateLimitExceededError: description: Rate limit exceeded headers: X-RateLimit-Limit: $ref: '#/components/headers/X-RateLimit-Limit' X-RateLimit-Remaining: $ref: '#/components/headers/X-RateLimit-Remaining' X-RateLimit-Reset: $ref: '#/components/headers/X-RateLimit-Reset' content: application/json: schema: $ref: '#/components/schemas/Error' headers: X-RateLimit-Limit: description: Maximum requests per window schema: type: integer X-RateLimit-Remaining: description: Requests remaining in current window schema: type: integer X-RateLimit-Reset: description: Unix timestamp when the window resets schema: type: integer securitySchemes: bearerAuth: type: http scheme: bearer description: >- API key with dtk_live_ prefix. Obtain it from API Keys in the account sidebar. New keys start with all supported scopes selected for the chosen project, and you can narrow them to specific read/write scopes as needed. ```` This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.