> ## Documentation Index > Fetch the complete documentation index at: https://docs.app.strix.ai/llms.txt > Use this file to discover all available pages before exploring further. # List LLM requests > Self-hosted installs only. Lists the LLM provider requests that the engine made for the runs of the authenticated organization, newest first. Each entry is one provider attempt: a success, a refusal, a timeout or one retry. An entry holds the provider request ID, the status code, the model, the latency, the token usage, the response headers as the provider sent them and the free-form provider details. An entry never holds a prompt or a completion. Raw request and response bodies are never stored. Hosted installs return 404. ## OpenAPI ````yaml /openapi.json get /llm-requests openapi: 3.1.0 info: title: Strix API version: 1.0.0 description: >- Public REST API for the Strix autonomous penetration testing platform. Manage scans, vulnerabilities, assets, schedules, API tokens, and webhooks. servers: - url: /api/v1 description: Strix v1 API security: - BearerAuth: [] tags: - name: Scans description: Launch, monitor, and manage security scans. - name: Vulnerabilities description: View and triage discovered vulnerabilities. - name: Assets description: Domains and repositories registered for scanning. - name: Schedules description: Recurring scan schedules (Pro plan). - name: Tokens description: Manage API tokens for authentication. - name: Webhooks description: Configure webhook subscriptions for real-time event notifications. - name: Organization description: Workspace configuration for the authenticated organization. - name: Members description: Manage organization members and roles. - name: Invitations description: List and revoke organization invitations. - name: PR Reviews description: Automated security review of pull requests. - name: Connectors description: Network connectors for scanning internal/private targets. - name: Knowledge description: >- Organization knowledge base: documents, policies, and repo profiles that steer the agent. - name: Uploads description: Upload source/code/documentation archives for whitebox scans. - name: Integrations description: Third-party integrations (GitLab, Bitbucket, ticketing). - name: Chat description: Conversational agent sessions. - name: Analytics description: Aggregate dashboard analytics. - name: Test Users description: >- Per-domain test accounts (with optional MFA) the agent authenticates as during scans. - name: License description: Self-hosted license state, entitlements, and aggregate usage. - name: Supply Chain description: >- SBOM inventory, supply-chain findings, scans, and policy for connected repositories. - name: CLI description: >- Device authorization endpoints that let the Strix CLI and coding agents sign in and receive an API token. - name: Billing description: Credit balance, agent-payable top-ups, and automatic top-up settings. - name: Workspaces description: List, create, and switch workspaces. - name: Asset Discovery description: >- Passive reconnaissance runs that map the organization's external attack surface, and the discovered-asset inventory they fill. - name: Containers description: >- Container registries, tracked images, image scans, findings, and SBOM components. - name: Run Logs description: Persisted engine logs and the diagnostics bundle of a self-hosted install. paths: /llm-requests: get: tags: - Run Logs summary: List LLM requests description: >- Self-hosted installs only. Lists the LLM provider requests that the engine made for the runs of the authenticated organization, newest first. Each entry is one provider attempt: a success, a refusal, a timeout or one retry. An entry holds the provider request ID, the status code, the model, the latency, the token usage, the response headers as the provider sent them and the free-form provider details. An entry never holds a prompt or a completion. Raw request and response bodies are never stored. Hosted installs return 404. operationId: listLlmRequests parameters: - name: id in: query description: Return only the attempt with this ID. schema: type: string format: uuid - name: q in: query description: >- Match the provider request ID, response ID, call ID, run ID, run name, model, provider, agent name, error type, error message, or a value in the response headers or details. schema: type: string maxLength: 200 - name: outcome in: query description: Return only attempts with this outcome. schema: type: string enum: - success - error - name: run_type in: query description: >- Return only attempts of this run type. `app` is a dashboard call such as a generated title or a compliance narrative. `connection_test` is Settings → LLM Models → Test connection. schema: type: string enum: - scan - pr_review - chat - auth_check - asset_discovery - app - connection_test - name: run_id in: query description: Return only attempts of this run. schema: type: string - name: provider in: query description: Return only attempts to this provider. schema: type: string - name: model in: query description: Return only attempts with this model. schema: type: string - name: limit in: query description: Maximum number of attempts to return. schema: type: integer minimum: 1 maximum: 200 default: 50 - name: offset in: query description: >- Number of attempts to skip before the first returned attempt. Ignored when `before` is set. schema: type: integer minimum: 0 default: 0 - name: before in: query description: >- ID of the last attempt of the previous page (`next_before`). The page continues after that attempt, so attempts that arrive between pages do not shift or repeat the results. schema: type: string format: uuid responses: '200': description: LLM provider attempts of the organization. content: application/json: schema: type: object properties: entries: type: array items: $ref: '#/components/schemas/LlmRequestEntry' total: type: integer description: Attempts that match the filters. total_unfiltered: type: integer description: Attempts of the organization before any filter. providers: type: array items: type: string description: Distinct providers of the organization, for filter menus. models: type: array items: type: string description: Distinct models of the organization, for filter menus. retention_days: type: integer description: >- Days an attempt stays before the retention cron removes it (LLM_REQUEST_LOG_RETENTION_DAYS). limit: type: integer offset: type: integer has_more: type: boolean description: True when another page exists after the last entry. next_before: type: - string - 'null' format: uuid description: >- Value for the `before` parameter of the next page. Null on the last page. required: - entries - total - total_unfiltered - retention_days - limit - offset - has_more - next_before '400': description: >- `outcome` or `run_type` is not one of the listed values, or `id` or `before` is not a UUID. '401': $ref: '#/components/responses/Unauthorized' '403': description: The caller is not an unrestricted organization admin. '404': description: LLM requests are not available on hosted installs. security: - BearerAuth: - logs:read components: schemas: LlmRequestEntry: type: object description: >- One LLM provider attempt that the engine made. Holds provider metadata only. properties: id: type: string format: uuid run_type: type: - string - 'null' enum: - scan - pr_review - chat - auth_check - asset_discovery - app - connection_test - null run_id: type: - string - 'null' run_name: type: - string - 'null' description: The engine run name, which also names the run log. run_href: type: - string - 'null' description: Dashboard path of the run, when the run has a page. call_id: type: string description: The engine call ID, unique per attempt. route: type: string enum: - litellm - openai provider: type: - string - 'null' description: For example `anthropic`. model: type: string api_host: type: - string - 'null' description: Host of the provider endpoint. streaming: type: boolean outcome: type: string enum: - success - error status_code: type: - integer - 'null' description: HTTP status from the provider. provider_request_id: type: - string - 'null' description: >- The request ID that the provider assigned. Quote this ID to the provider support team. response_id: type: - string - 'null' description: The response ID from the provider body. error_type: type: - string - 'null' description: Exception class of a failed attempt. error_message: type: - string - 'null' description: >- Error text of a failed attempt. The engine redacts secrets before it records the text. started_at: type: string format: date-time finished_at: type: string format: date-time duration_ms: type: integer input_tokens: type: - integer - 'null' output_tokens: type: - integer - 'null' cached_input_tokens: type: - integer - 'null' total_tokens: type: - integer - 'null' cost_usd: type: - number - 'null' agent_id: type: - string - 'null' agent_name: type: - string - 'null' retry_attempt: type: integer description: >- 0 for the first attempt of a turn, then 1, 2 and so on for each retry. request_bytes: type: - integer - 'null' description: >- Size in bytes of the JSON request body that the SDK sent. Not the compressed size on the wire. response_bytes: type: - integer - 'null' description: >- Size in bytes of the JSON response body, or of the error body of a failed attempt. time_to_first_token_ms: type: - integer - 'null' description: >- Streaming attempts only. Time from the start of the request to the first stream event. finish_reason: type: - string - 'null' description: >- The stop reason from the provider, for example `stop`, `tool_calls`, `max_tokens` or `incomplete:max_output_tokens`. response_headers: type: - object - 'null' additionalProperties: type: string description: >- Response headers from the provider, by lower-case header name. Headers whose name looks like a credential (authorization, cookie, api-key, secret, signature, session) are removed. Values are redacted and cut at 512 characters. At most 64 headers are kept. details: type: - object - 'null' description: >- Free-form provider and SDK metadata beyond the typed fields, for example the raw usage object, the stop sequence, the service tier or the system fingerprint. Content-shaped and secret-shaped keys are removed before storage. Limited to 16 KiB. required: - id - call_id - route - model - streaming - outcome - started_at - finished_at - duration_ms - retry_attempt ErrorResponse: type: object properties: detail: type: string description: Human-readable explanation of the error. code: type: string description: >- Stable machine-readable error code. `insufficient_scope` means the token does not hold the scope that this endpoint requires. required_scope: $ref: '#/components/schemas/ApiV1Scope' description: >- Scope that the caller must add to the token. Returned with the `insufficient_scope` code. docs: type: string format: uri description: Documentation page that explains how to resolve the error. hint: type: string description: >- One instruction that resolves the error. For `insufficient_scope`, a CLI session gets the `strix cloud session scopes set full` or `strix cloud login --scope-profile full` command, and an API token gets the settings page where the user creates a token with the scope. When the owner's role cannot hold the scope, the hint asks for a role change instead. required: - detail ApiV1Scope: type: string enum: - scans:read - scans:write - vulnerabilities:read - vulnerabilities:write - dependencies:read - schedules:read - schedules:write - assets:read - assets:write - organizations:read - organizations:write - members:read - members:write - invitations:read - invitations:write - webhooks:read - webhooks:write - tokens:write - audit:read - pr_reviews:read - pr_reviews:write - connectors:read - connectors:write - knowledge:read - knowledge:write - uploads:write - integrations:read - integrations:write - chat:read - chat:write - scans:message - analytics:read - llm:read - llm:write - test_users:read - test_users:write - discovery:read - discovery:write - license:read - supply_chain:read - supply_chain:write - billing:read - billing:write - containers:read - containers:write - logs:read responses: Unauthorized: description: Missing or invalid API token. content: application/json: schema: $ref: '#/components/schemas/ErrorResponse' securitySchemes: BearerAuth: type: http scheme: bearer description: >- API token obtained from the Tokens endpoint or CLI device login. Include as `Authorization: Bearer `. Requests made with a managed CLI session also include `X-Strix-Workspace: ` to pin a process to the workspace it started in; recovery endpoints report the current workspace after a concurrent switch. ````