# Models: C > API models starting with C. ## `Call` - `run_id: str` (`runId` on the wire) - `node_id: str | None` (`nodeId` on the wire) (optional) - `scope: str | None` (optional) - `tool_id: str` (`toolId` on the wire) - `arguments: Any | None` (optional) - `result: Any | None` (optional) - `from_event: Optional[Literal[True]]` (`fromEvent` on the wire) (optional) ## `CallDifferenceCount` One way a run's calls differed from a correction's expected decisions; it can quote argument values. - `tool: str` - `kind: Literal['count']` - `expected: int` - `actual: int` - `refused: int | None` (optional) ## `CallDifferenceFieldValue` One way a run's calls differed from a correction's expected decisions; it can quote argument values. - `tool: str` - `kind: Literal['field-value']` - `field: str` - `expected: Any` - `actual: Any` - `refused: bool | None` (optional) ## `CallDifferenceMissingField` One way a run's calls differed from a correction's expected decisions; it can quote argument values. - `tool: str` - `kind: Literal['missing-field']` - `field: str` - `expected: Any` - `refused: bool | None` (optional) ## `CallDifferenceMustNotCall` One way a run's calls differed from a correction's expected decisions; it can quote argument values. - `tool: str` - `kind: Literal['must-not-call']` - `actual: int` - `refused: int | None` (optional) ## `CallUsage` - `usage: kindgi.client.ModelCallTokens` - `cost_usd: float | None` (`costUsd` on the wire) (optional) - `duration_ms: int | None` (`durationMs` on the wire) (optional) - `served_model: str | None` (`servedModel` on the wire) (optional) ## `CandidateItem` An output's score: Σ yesWeight and Σ totalWeight over its corrected items, and over those among the first `k` ranked items. - `yes_weight: float` (`yesWeight` on the wire) - `total_weight: float` (`totalWeight` on the wire) - `items: int` - `corrected_items: int` (`correctedItems` on the wire) - `top_k: kindgi.client.TopK` (`topK` on the wire) - `fresh: kindgi.client.Fresh | None` (optional) ## `Capability` - `root: str` ## `Capability4` Capability declaration — see `@kindgi/capabilities`. Additional properties are permitted so new capability kinds do not require a wire change. ## `Capability5` - `root: str` ## `CapabilityCollectionPage` - `data: list[kindgi.client.CapabilityDescriptor]` - `next_cursor: str | None` (`nextCursor` on the wire) (optional) - `has_more: bool` (`hasMore` on the wire) ## `CapabilityDescriptor` - `id: str` - `feature: Union[str, Annotated[Union[Literal['structured-output', 'vision', 'audio-input', 'audio-output', 'tool-use', 'parallel-tool-use', 'thinking', 'long-context', 'code-execution', 'web-search', 'file-search', 'streaming', 'batch'], str], ]]` - `description: str` - `kind: str | None` (optional) - `params_schema: dict[str, Any] | None` (`paramsSchema` on the wire) (optional) - `providers: list[kindgi.client.CapabilityProvider] | None` (optional) ## `CapabilityProvider` A provider of the tenant with a model that has the feature. - `provider_id: str` (`providerId` on the wire) - `models: list[str]` ## `Cases` The cases the comparison graded: each correction it checked is a case, and it passes when every repetition did. - `min_passing: int | None` (`minPassing` on the wire) (optional) - `no_new_failures: bool | None` (`noNewFailures` on the wire) (optional) ## `ChangeCorrectionBody` A re-sort, a status, or both (the re-sort first). `fixed` is set only by a pin or a deploy, and `removed` by `POST /v1/corrections/{correctionId}/remove`. - `sorted_to: Optional[Literal['improve', 'developer', 'owner']]` (`sortedTo` on the wire) (optional) - `status: Optional[Literal['open', 'being-fixed', 'wont-fix']]` (optional) - `link: str | None` (optional) - `reason: str | None` (optional) ## `Changes` The first repetition's items against the corrected ones. - `kept: list[kindgi.client.KeptItem]` - `dropped: list[kindgi.client.DroppedItem]` - `new: list[kindgi.client.NewItem]` ## `ClaimMappingScopesSpec` Claim path + optional delimiter for extracting Kindgi OAuth scopes from provider claims. When absent, scopes come from the token response `scope` field instead. - `claim: str` - `delimiter: str | None` (optional) ## `ClaimMappingSpec` Per-provider claim mapping. All fields optional; closer to OIDC defaults (sub/email/name) → fewer overrides needed. - `user_id: str | None` (`userId` on the wire) (optional) - `email: str | None` (optional) - `display_name: str | None` (`displayName` on the wire) (optional) - `scopes: kindgi.client.ClaimMappingScopesSpec | None` (optional) - `metadata: list[kindgi.client.Metadatum] | None` (optional) ## `ClaimOrchestratorRoundBody` - `claimant: str` - `hold_seconds: int` (`holdSeconds` on the wire) ## `CodeArtifactRefFilesystem` Handler-artifact pointer. Discriminated on `kind`: `oci` is the deploy-pipeline shape (image + module path + artifactVersion); `filesystem` is the local-development shape — absolute host path at the pack's on-disk handler file. Production servers SHOULD reject `filesystem`. - `kind: Literal['filesystem']` - `module_path: str` (`modulePath` on the wire) ## `CodeArtifactRefOci` Handler-artifact pointer. Discriminated on `kind`: `oci` is the deploy-pipeline shape (image + module path + artifactVersion); `filesystem` is the local-development shape — absolute host path at the pack's on-disk handler file. Production servers SHOULD reject `filesystem`. - `kind: Literal['oci']` - `image_ref: str` (`imageRef` on the wire) - `module_path: str` (`modulePath` on the wire) - `artifact_version: str` (`artifactVersion` on the wire) ## `CodeArtifactRefUnrecognized` A `kind` this client doesn't know yet, from a newer runtime: the object as it came, every field kept. - `kind: str` ## `Comparison` - `max_age_hours: float | None` (`maxAgeHours` on the wire) (optional) - `test_set: kindgi.client.ReleaseGateSpecComparisonTestSet | None` (`testSet` on the wire) (optional) ## `Comparison1` - `test_run_id: uuid.UUID | None` (`testRunId` on the wire) (optional) - `role: Union[Literal['reference', 'candidate', 'proof'], str]` - `part: Union[Literal['search', 'hold-out', 'all'], str]` - `block_id: str | None` (`blockId` on the wire) (optional) - `changed: dict[str, Any] | None` (optional) - `score: float | None` (optional) - `failed: str | None` (optional) - `refused: list[kindgi.client.RefusedItem] | None` (optional) - `why: str | None` (optional) ## `ComparisonCandidateAgent` What ran on the cases: an agent version, or a flow version (with any versions it swapped in). - `kind: Literal['agent']` - `agent_id: str` (`agentId` on the wire) - `version: str` - `pins_digest: str | None` (`pinsDigest` on the wire) (optional) - `overrides: kindgi.client.Overrides | None` (optional) ## `ComparisonCandidateFlow` What ran on the cases: an agent version, or a flow version (with any versions it swapped in). - `kind: Literal['flow']` - `flow_id: str` (`flowId` on the wire) - `version: str` - `versions: kindgi.client.FlowVersionOverrides | None` (optional) ## `ComparisonCaseResult` One case of a comparison: its replay runs, the scores, the items kept, dropped and new, the tool calls, and why it didn't run when it didn't. - `case_id: str` (`caseId` on the wire) - `run_ids: list[str]` (`runIds` on the wire) - `baseline: kindgi.client.ComparisonCaseResultBaseline` - `candidate: list[kindgi.client.CandidateItem]` - `changes: kindgi.client.Changes | None` (optional) - `tools: list[kindgi.client.ComparisonCaseResultTool] | None` (optional) - `corrections: list[kindgi.client.ComparisonCaseResultCorrection] | None` (optional) - `correction_differences: list[kindgi.client.CorrectionDifference] | None` (`correctionDifferences` on the wire) (optional) - `cost_usd: list[kindgi.client.CostUsdItem] | None` (`costUsd` on the wire) (optional) - `unchecked: str | None` (optional) - `diverged: bool` - `refused_writes: int` (`refusedWrites` on the wire) - `no_context: bool` (`noContext` on the wire) - `approval_skipped: bool` (`approvalSkipped` on the wire) - `error: str | None` (optional) - `stopped: kindgi.client.Stopped | None` (optional) - `rescored: Optional[Literal[False]]` (optional) ## `ComparisonCaseResultBaseline` An output's score: Σ yesWeight and Σ totalWeight over its corrected items, and over those among the first `k` ranked items. - `yes_weight: float` (`yesWeight` on the wire) - `total_weight: float` (`totalWeight` on the wire) - `items: int` - `corrected_items: int` (`correctedItems` on the wire) - `top_k: kindgi.client.TopK` (`topK` on the wire) - `fresh: kindgi.client.Fresh | None` (optional) ## `ComparisonCaseResultCorrection` - `correction_id: str` (`correctionId` on the wire) - `pass_: bool` (`pass` on the wire) - `repetitions_passed: int` (`repetitionsPassed` on the wire) - `repetitions: int` - `why: str | None` (optional) - `parts: kindgi.client.Parts | None` (optional) ## `ComparisonCaseResultTool` - `step: int` - `call_id: str` (`callId` on the wire) - `tool_id: str` (`toolId` on the wire) - `tool_version: str` (`toolVersion` on the wire) - `arguments: Any` - `source: Literal['live', 'recorded', 'refused']` - `recomputed: bool | None` (optional) - `reason: str | None` (optional) ## `ComparisonMetric` One metric, the recorded runs beside the candidate. `null` where a side had no evidence from corrections: no corrected items, and for `passRate` no graded `expect` either. `n` / `weight` are the candidate's evidence, the cases it rests on and the weight behind them. For `passRate`, those are the cases with corrected items or a graded `expect`, and each graded part of an `expect`, its decisions and its answer, is one check of weight 1 beside the items' correction weight. For `caseCoverage` and `weightedPrecisionAtK`, they're the cases with corrected items only. `baselineN` / `baselineWeight` are the recorded side's. - `baseline: float | None` - `candidate: float | None` - `delta: float | None` - `n: int` - `weight: float` - `baseline_n: int` (`baselineN` on the wire) - `baseline_weight: float` (`baselineWeight` on the wire) - `direction: Literal['higher']` - `k: int | None` (optional) - `spread: float | None` (optional) - `fresh_weight: float | None` (`freshWeight` on the wire) (optional) ## `CompleteApprovalBody` - `decision: Literal['approve', 'reject', 'escalate', 'withdraw']` - `rationale: str | None` (optional) - `value: Any | None` (optional) - `fact_text: str | None` (`factText` on the wire) (optional) ## `CompleteApprovalResult` A recorded decision, and what it did to the run waiting on the approval. `runStatus` is the run's status when the decision couldn't resolve its waitpoint because the run had already ended (e.g. `cancelled` after the decision was recorded); the decision stands. - `kind: Union[Literal['terminal', 'escalated'], str]` - `approval: kindgi.client.Approval` - `decision: kindgi.client.ReviewDecision` - `next_approval: kindgi.client.Approval | None` (`nextApproval` on the wire) (optional) - `waitpoint_resolved: bool` (`waitpointResolved` on the wire) - `run_status: Optional[Annotated[Union[Literal['pending', 'running', 'suspended', 'completed', 'failed', 'cancelled'], str], ]]` (`runStatus` on the wire) (optional) - `resume: Optional[Annotated[kindgi.client.CompleteApprovalResultResumeOk | kindgi.client.CompleteApprovalResultResumeFailed | kindgi.client.CompleteApprovalResultResumeUnrecognized, ]]` (optional) ## `CompleteApprovalResultResumeFailed` How the run went on, when this call resumed it (the runtime resumes inline): `ok`, or `failed` with the run's error, e.g. `tool-version-unresolvable` when a tool version the turn started with is gone, or `rate-limit-exceeded` when a budget kept per tenant refused work the turn needed. The decision stands either way. - `kind: Literal['failed']` - `code: str` - `message: str` ## `CompleteApprovalResultResumeOk` How the run went on, when this call resumed it (the runtime resumes inline): `ok`, or `failed` with the run's error, e.g. `tool-version-unresolvable` when a tool version the turn started with is gone, or `rate-limit-exceeded` when a budget kept per tenant refused work the turn needed. The decision stands either way. - `kind: Literal['ok']` ## `CompleteApprovalResultResumeUnrecognized` A `kind` this client doesn't know yet, from a newer runtime: the object as it came, every field kept. - `kind: str` ## `ComplianceEvidence` One evidence record — matches `@kindgi/specs/compliance-evidence.schema.json` (`payload` is a versioned document opaque to the wire schema). - `id: str` - `tenant_id: uuid.UUID` (`tenantId` on the wire) - `project_id: uuid.UUID | None` (`projectId` on the wire) (optional) - `kind: str` - `timestamp: AwareDatetime` - `actor: kindgi.client.Actor | None` (optional) - `subject: kindgi.client.Subject1 | None` (optional) - `outcome: Optional[Annotated[Union[Literal['allowed', 'denied', 'succeeded', 'failed', 'escalated'], str], ]]` (optional) - `payload: dict[str, Any]` - `provenance_ref: kindgi.client.ProvenanceRef2 | None` (`provenanceRef` on the wire) (optional) - `signature: kindgi.client.Signature1 | None` (optional) ## `ComplianceEvidenceCollectionPage` - `data: list[kindgi.client.ComplianceEvidence]` - `has_more: bool` (`hasMore` on the wire) - `next_cursor: str | None` (`nextCursor` on the wire) (optional) ## `ConnectFinish` - `kind: Union[Literal['connected', 'declined', 'pick-installation'], str]` - `request: kindgi.client.ConnectRequest` - `connection: kindgi.client.Connection | None` (optional) - `reviewer_seat: bool | None` (`reviewerSeat` on the wire) (optional) ## `ConnectRequest` A connect: the owner goes to `connectUrl`, consents at the app, and comes back. A run that needs a connection it lacks waits on one. - `request_id: uuid.UUID` (`requestId` on the wire) - `env_name: str` (`envName` on the wire) - `app_id: uuid.UUID` (`appId` on the wire) - `app: str` - `owner: kindgi.client.ConnectionOwner` - `permissions: list[str]` - `says: list[str]` - `reconnects: uuid.UUID | None` (optional) - `run_id: str | None` (`runId` on the wire) (optional) - `name: str | None` (optional) - `started_by: str | None` (`startedBy` on the wire) (optional) - `state: Union[Literal['pending', 'connected', 'declined', 'expired', 'failed'], str]` - `failure: str | None` (optional) - `missing: list[str] | None` (optional) - `installations: list[kindgi.client.ConnectRequestInstallation] | None` (optional) - `connection_id: uuid.UUID | None` (`connectionId` on the wire) (optional) - `connect_url: AnyUrl` (`connectUrl` on the wire) - `expires_at: AwareDatetime` (`expiresAt` on the wire) - `created_at: AwareDatetime` (`createdAt` on the wire) - `created_by: str` (`createdBy` on the wire) ## `ConnectRequestInstallation` A GitHub App installation the person can reach. - `installation_id: str` (`installationId` on the wire) - `account: kindgi.client.ConnectionAccount` ## `ConnectStart` - `authorize_url: AnyUrl` (`authorizeUrl` on the wire) ## `Connection` One owner's account at an app, through a registration, in one env. Never its token. - `connection_id: uuid.UUID` (`connectionId` on the wire) - `env_name: str` (`envName` on the wire) - `app_id: uuid.UUID` (`appId` on the wire) - `app: str` - `owner: kindgi.client.ConnectionOwner` - `account: kindgi.client.ConnectionAccount` - `permissions: list[str]` - `state: Union[Literal['working', 'needs-reconnecting', 'removing', 'removed'], str]` - `state_reason: str | None` (`stateReason` on the wire) (optional) - `installation_id: str | None` (`installationId` on the wire) (optional) - `access_ends_at: AwareDatetime | None` (`accessEndsAt` on the wire) (optional) - `connected_at: AwareDatetime` (`connectedAt` on the wire) - `last_used_at: AwareDatetime | None` (`lastUsedAt` on the wire) (optional) - `created_at: AwareDatetime` (`createdAt` on the wire) - `created_by: str` (`createdBy` on the wire) - `updated_at: AwareDatetime` (`updatedAt` on the wire) - `removed_at: AwareDatetime | None` (`removedAt` on the wire) (optional) - `removed_by: str | None` (`removedBy` on the wire) (optional) - `removed_reason: Optional[Annotated[Union[Literal['disconnected', 'revoked-at-app', 'uninstalled', 'registration-removed', 'replaced'], str], ]]` (`removedReason` on the wire) (optional) - `withdrawn: Optional[Annotated[Union[Literal['at-app', 'app-failed', 'kept-for-other-connection', 'not-supported'], str], ]]` (optional) ## `ConnectionAccount` Whose account a connection is at the app. - `id: str` - `label: str` ## `ConnectionApp` A tenant's registration of its own client at an app. Secrets are references, never values. - `app_id: uuid.UUID` (`appId` on the wire) - `app: str` - `label: str | None` (optional) - `oauth2: kindgi.client.ConnectionOAuth2App | None` (optional) - `github_app: kindgi.client.ConnectionGitHubApp | None` (`githubApp` on the wire) (optional) - `projects: list[str] | None` (optional) - `redirect_url: AnyUrl` (`redirectUrl` on the wire) - `notices_url: AnyUrl` (`noticesUrl` on the wire) - `created_at: AwareDatetime` (`createdAt` on the wire) - `updated_at: AwareDatetime` (`updatedAt` on the wire) - `removed_at: AwareDatetime | None` (`removedAt` on the wire) (optional) ## `ConnectionAppPage` - `data: list[kindgi.client.ConnectionApp]` - `next_cursor: str | None` (`nextCursor` on the wire) (optional) - `has_more: bool` (`hasMore` on the wire) ## `ConnectionAppRemoved` - `connections_removed: int` (`connectionsRemoved` on the wire) ## `ConnectionEntry` An app this server can connect to. - `id: str` - `name: str` - `owners: list[Annotated[Union[Literal['member', 'company'], str], ]]` - `sign_in: kindgi.client.SignIn` (`signIn` on the wire) - `permissions: dict[str, kindgi.client.ConnectionEntryPermission]` - `removal: kindgi.client.Removal | None` (optional) - `setup_guide: str` (`setupGuide` on the wire) ## `ConnectionEntryList` - `data: list[kindgi.client.ConnectionEntry]` ## `ConnectionEntryPermission` - `says: str` - `changes: bool` ## `ConnectionEntrySignInTenant` Present when the sign-in happens in a directory of the app's own (Microsoft's tenant, not a Kindgi tenant). A registration's `oauth2.tenant` names it; without `default`, each registration must give its own. - `default: str | None` (optional) ## `ConnectionGitHubApp` - `app_number: int` (`appNumber` on the wire) - `slug: str` - `client_id: str` (`clientId` on the wire) - `client_secret: kindgi.client.ConnectionSecretRef` (`clientSecret` on the wire) - `private_key: kindgi.client.ConnectionSecretRef` (`privateKey` on the wire) - `webhook_secret: kindgi.client.ConnectionSecretRef` (`webhookSecret` on the wire) ## `ConnectionNeed` - `app: str` - `permissions: list[kindgi.client.Permission]` - `as_: Union[Literal['person', 'company'], str]` (`as` on the wire) ## `ConnectionOAuth2App` - `client_id: str` (`clientId` on the wire) - `client_secret: kindgi.client.ConnectionSecretRef` (`clientSecret` on the wire) - `tenant: str | None` (optional) ## `ConnectionOwner` Who owns a connection: one member (`kind: member`, with their `userId`), or the company (`kind: company`, the tenant's own account). - `kind: Union[Literal['member', 'company'], str]` - `user_id: str | None` (`userId` on the wire) (optional) ## `ConnectionPage` - `data: list[kindgi.client.Connection]` - `next_cursor: str | None` (`nextCursor` on the wire) (optional) - `has_more: bool` (`hasMore` on the wire) ## `ConnectionSecretRef` A secret by name in the deployment's secrets store, resolved at tenant scope. The registration keeps only this reference, never the value. - `env_name: str` (`envName` on the wire) - `name: str` ## `ConsoleLink` One of the deployment's own links, which the console shows on every page. - `label: str` - `url: AnyUrl` ## `Conversation` - `id: uuid.UUID` - `tenant_id: uuid.UUID` (`tenantId` on the wire) - `agent_id: str` (`agentId` on the wire) - `agent_version: str` (`agentVersion` on the wire) - `title: str` - `participant_id: str | None` (`participantId` on the wire) (optional) - `project_id: uuid.UUID | None` (`projectId` on the wire) (optional) - `opened_by: str | None` (`openedBy` on the wire) (optional) - `scope: dict[str, Any]` - `status: Union[Literal['open', 'closed'], str]` - `opened_at: AwareDatetime` (`openedAt` on the wire) - `closed_at: AwareDatetime | None` (`closedAt` on the wire) (optional) - `unregistered_at: AwareDatetime | None` (`unregisteredAt` on the wire) (optional) - `turn_count: int` (`turnCount` on the wire) - `last_message_at: AwareDatetime | None` (`lastMessageAt` on the wire) (optional) - `metadata: dict[str, Any] | None` (optional) ## `ConversationCollectionPage` - `data: list[kindgi.client.Conversation]` - `next_cursor: str | None` (`nextCursor` on the wire) (optional) - `has_more: bool` (`hasMore` on the wire) - `opened_by: Optional[Annotated[Union[Literal['me'], str], ]]` (`openedBy` on the wire) (optional) ## `ConversationMessage` - `sequence: int` - `role: Union[Literal['user', 'agent', 'tool', 'system'], str]` - `content: Any` - `files: list[kindgi.client.FileValue] | None` (optional) - `tool_call: kindgi.client.ToolCall | None` (`toolCall` on the wire) (optional) - `actor: str | None` (optional) - `created_at: AwareDatetime` (`createdAt` on the wire) ## `ConversationMessageCollectionPage` - `data: list[kindgi.client.ConversationMessage]` - `next_cursor: str | None` (`nextCursor` on the wire) (optional) - `has_more: bool` (`hasMore` on the wire) ## `ConversationPolicy` - `history_limit: int | None` (`historyLimit` on the wire) (optional) - `hitl_after_turns: int | None` (`hitlAfterTurns` on the wire) (optional) ## `ConversationStatus` - `root: Literal['open', 'closed']` ## `Corrected` What people said about this item on the replay itself, once they corrected it there. - `yes_weight: float` (`yesWeight` on the wire) - `total_weight: float` (`totalWeight` on the wire) ## `CorrectedAnswerItem` - `key: str` - `pointer: str | None` (optional) - `rank: int | None` (optional) - `verdict: Union[Literal['right', 'wrong'], str]` ## `CorrectedAnswerItemInput` - `key: str` - `pointer: str | None` (optional) - `rank: int | None` (optional) - `verdict: Literal['right', 'wrong']` ## `CorrectedComparisonCost` What the candidate's replays cost in model calls, as the runtime reported it. Absent from a summary when no case's replays reported a cost. - `candidate: float` - `spread: float | None` (optional) - `cases: int` ## `CorrectedComparisonResult` A comparison's `result` (a `corrections` test run's): the summary and each case. `TestRun.result` stays an open object, since each kind has its own; the clients read it as this (TS `comparisonOf(run)`, Python `comparison_of(run)`). - `summary: kindgi.client.CorrectedComparisonSummary` - `per_case: list[kindgi.client.ComparisonCaseResult]` (`perCase` on the wire) ## `CorrectedComparisonSummary` What a comparison concluded: the candidate beside the recorded runs, the case counts, and the metrics. What a release gate reads. - `test_run_id: str` (`testRunId` on the wire) - `status: Literal['completed', 'partial', 'failed']` - `completed_at: AwareDatetime` (`completedAt` on the wire) - `test_set: kindgi.client.CorrectedComparisonSummaryTestSet` (`testSet` on the wire) - `candidate: kindgi.client.ComparisonCandidateAgent | kindgi.client.ComparisonCandidateFlow` - `baseline: kindgi.client.CorrectedComparisonSummaryBaselineRecorded | kindgi.client.CorrectedComparisonSummaryBaselineVersion` - `scope: kindgi.client.CorrectedComparisonSummaryScope` - `cases: int` - `diverged: int` - `refused_writes: int` (`refusedWrites` on the wire) - `errors: int` - `erased: int | None` (optional) - `rescore_of: str | None` (`rescoreOf` on the wire) (optional) - `not_rescored: int | None` (`notRescored` on the wire) (optional) - `stopped: int` - `reads: Literal['recorded', 'live']` - `group_weights: Optional[Literal['as-recorded', 'experts-only']]` (`groupWeights` on the wire) (optional) - `sample: kindgi.client.TestRunSample | None` (optional) - `sampling: kindgi.client.Sampling` - `repetitions: int` - `metrics: kindgi.client.Metrics` - `corrections: kindgi.client.Corrections | None` (optional) - `cost_usd: kindgi.client.CorrectedComparisonCost | None` (`costUsd` on the wire) (optional) - `unchecked: kindgi.client.CorrectedComparisonUnchecked | None` (optional) ## `CorrectedComparisonSummaryBaselineRecorded` What the candidate was compared with: `recorded` (the test set's recorded runs, with the versions that served them), or another version. - `kind: Literal['recorded']` - `versions: list[kindgi.client.Version]` ## `CorrectedComparisonSummaryBaselineVersion` What the candidate was compared with: `recorded` (the test set's recorded runs, with the versions that served them), or another version. - `kind: Literal['version']` - `agent_id: str` (`agentId` on the wire) - `version: str` - `via: Literal['explicit', 'pinned']` - `pinned_scope: dict[str, Any] | None` (`pinnedScope` on the wire) (optional) ## `CorrectedComparisonSummarySamplingModel` - `provider_id: str` (`providerId` on the wire) - `model: str` - `runs: int` ## `CorrectedComparisonSummaryScope` Where the test set's corrections came from. - `project_id: str | None` (`projectId` on the wire) (optional) ## `CorrectedComparisonSummaryTestSet` - `id: str` - `version: str` ## `CorrectedComparisonUnchecked` The cases whose `expect` couldn't be graded, so they add no check to `passRate`: when `passRate` has no value, this says why. Absent from a summary when every case's was graded, and from older servers. - `cases: int` - `reasons: list[str]` ## `CorrectedItem` The corrected item of a run's output. - `key: str` - `pointer: str | None` (optional) - `rank: int | None` (optional) ## `CorrectedItemSummary` The corrections of one item of a case's output, summed up. - `key: str` - `pointer: str | None` (optional) - `rank: int | None` (optional) - `yes: int` - `no: int` - `yes_weight: float` (`yesWeight` on the wire) - `total_weight: float` (`totalWeight` on the wire) - `expert: kindgi.client.Expert | None` (optional) - `correction_ids: list[str] | None` (`correctionIds` on the wire) (optional) - `reasons: list[kindgi.client.Reason]` ## `CorrectedRunContext` What a corrected run needs besides its input to be replayed, captured when it was first corrected. For an agent turn: the conversation before it, what its retrievals returned, and the decision at its session approval gate. For a flow run: its tool calls with their results. For both: the env values its tools were sent. - `history: list[Any] | None` (optional) - `history_truncated: bool | None` (`historyTruncated` on the wire) (optional) - `retrieved: Any | None` (optional) - `recalled: Any | None` (optional) - `session_approval: kindgi.client.SessionApproval | None` (`sessionApproval` on the wire) (optional) - `flow: kindgi.client.CorrectedRunContextFlow | None` (optional) - `tool_env: dict[str, dict[str, str]] | None` (`toolEnv` on the wire) (optional) - `replay_of: str | None` (`replayOf` on the wire) (optional) ## `CorrectedRunContextFlow` For a flow run: what it did, kept at its first correction so it can be replayed. Every tool call it made with its result (at its tool nodes, in its agent steps' turns and in its sub-flows), at most 500, and its agent steps. - `calls: list[kindgi.client.Call]` - `steps: list[kindgi.client.Step]` - `truncated: bool | None` (optional) ## `CorrectedRunCopy` The stored copy of a corrected run's input and output, taken when it was first corrected. - `run_id: str` (`runId` on the wire) - `subject: kindgi.client.CorrectedSubject` - `input: Any` - `context: kindgi.client.CorrectedRunContext | None` (optional) - `output: Any` - `segments: list[kindgi.client.ScopeSegment] | None` (optional) - `captured_at: AwareDatetime` (`capturedAt` on the wire) ## `CorrectedSubject` What a corrected run ran: an agent at a version, or a flow at a version. - `kind: Union[Literal['agent', 'flow'], str]` - `id: str` - `version: str` ## `CorrectedTestCase` One case of a `corrections` test set: a copy of a corrected run with its items' corrections summed up. - `case_id: str` (`caseId` on the wire) - `run_id: str | None` (`runId` on the wire) (optional) - `subject: kindgi.client.CorrectedSubject` - `input: Any` - `context: kindgi.client.CorrectedRunContext | None` (optional) - `output: Any` - `segments: list[kindgi.client.ScopeSegment] | None` (optional) - `items: list[kindgi.client.CorrectedItemSummary]` - `corrections: list[kindgi.client.CorrectedTestCaseCorrection] | None` (optional) - `expect: kindgi.client.TestCaseExpect | None` (optional) - `expert: Optional[Literal[True]]` (optional) - `erased: Optional[Literal[True]]` (optional) ## `CorrectedTestCaseCollectionPage` - `data: list[kindgi.client.CorrectedTestCase]` - `next_cursor: str | None` (`nextCursor` on the wire) (optional) - `has_more: bool` (`hasMore` on the wire) ## `CorrectedTestCaseCorrection` - `correction_id: str` (`correctionId` on the wire) - `corrector_group_id: str | None` (`correctorGroupId` on the wire) (optional) - `expert: Optional[Literal[True]]` (optional) - `calls: list[kindgi.client.ExpectedCall]` - `not_called: list[kindgi.client.NotCalledItem]` (`notCalled` on the wire) ## `Correction` A person's correction of one run: whether its decisions (the tool calls) and its answer were right, and what was right instead. It says at least one of `decisions`, `answer` and `items`. - `id: str` - `tenant_id: str` (`tenantId` on the wire) - `project_id: str` (`projectId` on the wire) - `run_id: str` (`runId` on the wire) - `subject: kindgi.client.CorrectedSubject` - `decisions: Optional[Annotated[Union[Literal['right', 'wrong'], str], ]]` (optional) - `should_call: list[kindgi.client.ExpectedCall] | None` (`shouldCall` on the wire) (optional) - `must_not_call: list[str] | None` (`mustNotCall` on the wire) (optional) - `answer: Optional[Annotated[Union[Literal['right', 'wrong'], str], ]]` (optional) - `should_say: str | None` (`shouldSay` on the wire) (optional) - `why: str | None` (optional) - `items: list[kindgi.client.CorrectedAnswerItem] | None` (optional) - `corrector_group_id: str | None` (`correctorGroupId` on the wire) (optional) - `expert: Optional[Literal[True]]` (optional) - `made_by: kindgi.client.CorrectionMadeBy` (`madeBy` on the wire) - `participant_id: str | None` (`participantId` on the wire) (optional) - `status: Union[Literal['open', 'being-fixed', 'with-owner', 'fixed', 'wont-fix', 'removed'], str]` - `sorted_to: Optional[Annotated[Union[Literal['improve', 'developer', 'owner'], str], ]]` (`sortedTo` on the wire) (optional) - `sorted: kindgi.client.CorrectionSorted | None` (optional) - `link: Optional[Annotated[kindgi.client.CorrectionLink1 | kindgi.client.CorrectionLink2 | kindgi.client.CorrectionLinkUnrecognized, ]]` (optional) - `fixed_in: kindgi.client.CorrectionFixedIn | None` (`fixedIn` on the wire) (optional) - `reopened_in: kindgi.client.CorrectionReopenedIn | None` (`reopenedIn` on the wire) (optional) - `status_note: kindgi.client.CorrectionStatusNote | None` (`statusNote` on the wire) (optional) - `created_at: AwareDatetime` (`createdAt` on the wire) - `updated_at: AwareDatetime` (`updatedAt` on the wire) - `unregistered_at: AwareDatetime | None` (`unregisteredAt` on the wire) (optional) - `superseded_by: str | None` (`supersededBy` on the wire) (optional) ## `CorrectionCollectionPage` - `data: list[kindgi.client.Correction]` - `next_cursor: str | None` (`nextCursor` on the wire) (optional) - `has_more: bool` (`hasMore` on the wire) ## `CorrectionCounts` How many of a project's (or an agent's) live corrections are in each status. - `open: int` - `being_fixed: int` (`beingFixed` on the wire) - `with_owner: int` (`withOwner` on the wire) - `fixed: int` - `wont_fix: int` (`wontFix` on the wire) - `removed: int` ## `CorrectionDifference` - `correction_id: str` (`correctionId` on the wire) - `repetition: int` - `differences: list[kindgi.client.CallDifferenceCount | kindgi.client.CallDifferenceMustNotCall | kindgi.client.CallDifferenceMissingField | kindgi.client.CallDifferenceFieldValue]` ## `CorrectionFixedIn` - `version: str | None` (optional) - `how: Union[Literal['pin', 'latest', 'recorded-right'], str]` - `by: kindgi.client.CorrectionMadeBy | None` (optional) - `at: AwareDatetime` ## `CorrectionLink1` - `kind: Literal['improve-run', 'suggested-change']` - `id: str` ## `CorrectionLink2` - `kind: Literal['pull-request']` - `url: AnyUrl` ## `CorrectionLinkUnrecognized` A `kind` this client doesn't know yet, from a newer runtime: the object as it came, every field kept. - `kind: str` ## `CorrectionMadeBy` Who asserted a correction: the authenticated caller, never a typed name. - `kind: Union[Literal['user', 'service'], str]` - `id: str` ## `CorrectionReopenedIn` The version whose test run got a `fixed` correction wrong again. It holds no words: a client builds the sentence from the project's and the person's current names. - `version: str` - `how: Union[Literal['pin', 'went-back', 'unpin', 'latest'], str]` - `by: kindgi.client.CorrectionMadeBy | None` (optional) - `at: AwareDatetime` ## `CorrectionSorted` - `by: kindgi.client.By | kindgi.client.By1 | kindgi.client.CorrectionSortedByUnrecognized` - `reason: str | None` (optional) - `at: AwareDatetime` ## `CorrectionSortedByUnrecognized` A `kind` this client doesn't know yet, from a newer runtime: the object as it came, every field kept. - `kind: str` ## `CorrectionStatusNote` - `text: str` - `by: kindgi.client.CorrectionMadeBy` - `at: AwareDatetime` ## `CorrectionWithCopies` A correction with the stored copies of what was corrected. - `id: str` - `tenant_id: str` (`tenantId` on the wire) - `project_id: str` (`projectId` on the wire) - `run_id: str` (`runId` on the wire) - `subject: kindgi.client.CorrectedSubject` - `decisions: Optional[Annotated[Union[Literal['right', 'wrong'], str], ]]` (optional) - `should_call: list[kindgi.client.ExpectedCall] | None` (`shouldCall` on the wire) (optional) - `must_not_call: list[str] | None` (`mustNotCall` on the wire) (optional) - `answer: Optional[Annotated[Union[Literal['right', 'wrong'], str], ]]` (optional) - `should_say: str | None` (`shouldSay` on the wire) (optional) - `why: str | None` (optional) - `items: list[kindgi.client.CorrectedAnswerItem] | None` (optional) - `corrector_group_id: str | None` (`correctorGroupId` on the wire) (optional) - `expert: Optional[Literal[True]]` (optional) - `made_by: kindgi.client.CorrectionMadeBy` (`madeBy` on the wire) - `participant_id: str | None` (`participantId` on the wire) (optional) - `status: Union[Literal['open', 'being-fixed', 'with-owner', 'fixed', 'wont-fix', 'removed'], str]` - `sorted_to: Optional[Annotated[Union[Literal['improve', 'developer', 'owner'], str], ]]` (`sortedTo` on the wire) (optional) - `sorted: kindgi.client.CorrectionSorted | None` (optional) - `link: Optional[Annotated[kindgi.client.CorrectionLink1 | kindgi.client.CorrectionLink2 | kindgi.client.CorrectionLinkUnrecognized, ]]` (optional) - `fixed_in: kindgi.client.CorrectionFixedIn | None` (`fixedIn` on the wire) (optional) - `reopened_in: kindgi.client.CorrectionReopenedIn | None` (`reopenedIn` on the wire) (optional) - `status_note: kindgi.client.CorrectionStatusNote | None` (`statusNote` on the wire) (optional) - `created_at: AwareDatetime` (`createdAt` on the wire) - `updated_at: AwareDatetime` (`updatedAt` on the wire) - `unregistered_at: AwareDatetime | None` (`unregisteredAt` on the wire) (optional) - `superseded_by: str | None` (`supersededBy` on the wire) (optional) - `run: kindgi.client.CorrectedRunCopy` ## `Corrections` Corrections graded across the cases: shown for a release gate, not enforced. Absent when the test set carries no corrections. - `checked: int` - `passed: int` - `results: list[kindgi.client.Result] | None` (optional) ## `CorrectorGroup` - `id: str` - `tenant_id: str` (`tenantId` on the wire) - `scope: kindgi.client.CorrectorGroupScopeTenant | kindgi.client.CorrectorGroupScopeProject | kindgi.client.CorrectorGroupScopeAgent | kindgi.client.CorrectorGroupScopeUnrecognized` - `name: str` - `weight: float` - `description: str | None` (optional) - `assertable_by: kindgi.client.CorrectorGroupAssertableByView | None` (`assertableBy` on the wire) (optional) - `created_at: AwareDatetime` (`createdAt` on the wire) - `updated_at: AwareDatetime` (`updatedAt` on the wire) - `unregistered_at: AwareDatetime | None` (`unregisteredAt` on the wire) (optional) ## `CorrectorGroupAssertableBy` Who may assert the group: every part that is set must hold. A group with one is expert: a correction recorded while it is carries `expert`, and a comparison weighted `experts-only` counts only those. - `min_reviewer_role: Optional[Literal['standard', 'senior', 'admin']]` (`minReviewerRole` on the wire) (optional) - `principal_kinds: list[Literal['user', 'service']] | None` (`principalKinds` on the wire) (optional) - `principal_ids: list[kindgi.client.PrincipalId] | None` (`principalIds` on the wire) (optional) ## `CorrectorGroupAssertableByView` Who may assert the group, as the caller sees it: every part that is set must hold. `principalIds` is sent only to an admin on the group's scope; `principalCount` to everyone who reads it. - `min_reviewer_role: Optional[Annotated[Union[Literal['standard', 'senior', 'admin'], str], ]]` (`minReviewerRole` on the wire) (optional) - `principal_kinds: list[Annotated[Union[Literal['user', 'service'], str], ]] | None` (`principalKinds` on the wire) (optional) - `principal_ids: list[kindgi.client.PrincipalId] | None` (`principalIds` on the wire) (optional) - `principal_count: int | None` (`principalCount` on the wire) (optional) ## `CorrectorGroupCollectionPage` - `data: list[kindgi.client.CorrectorGroup]` - `next_cursor: str | None` (`nextCursor` on the wire) (optional) - `has_more: bool` (`hasMore` on the wire) ## `CorrectorGroupScopeAgent` Where a corrector group applies: the whole tenant, one project, or one agent in a project. - `kind: Literal['agent']` - `project_id: str` (`projectId` on the wire) - `agent_id: str` (`agentId` on the wire) ## `CorrectorGroupScopeProject` Where a corrector group applies: the whole tenant, one project, or one agent in a project. - `kind: Literal['project']` - `project_id: str` (`projectId` on the wire) ## `CorrectorGroupScopeTenant` Where a corrector group applies: the whole tenant, one project, or one agent in a project. - `kind: Literal['tenant']` ## `CorrectorGroupScopeUnrecognized` A `kind` this client doesn't know yet, from a newer runtime: the object as it came, every field kept. - `kind: str` ## `CostAggregateGroup` - `key: dict[str, str | None]` - `count: int` - `total_usd: float` (`totalUsd` on the wire) - `tokens: kindgi.client.CostTokenTotals` ## `CostAggregateResult` - `groups: list[kindgi.client.CostAggregateGroup]` - `total_groups: int | None` (`totalGroups` on the wire) (optional) - `truncated: bool | None` (optional) - `total_usd: float` (`totalUsd` on the wire) - `total_records: int` (`totalRecords` on the wire) - `tokens: kindgi.client.CostTokenTotals` - `time_range: kindgi.client.TimeRange` (`timeRange` on the wire) - `group_by: list[Annotated[Union[Literal['agentId', 'runId', 'category', 'providerId', 'day', 'month', 'tenant', 'conversationId', 'model', 'servedModel', 'projectId', 'orgId', 'rootRunId', 'flowId'], str], ]]` (`groupBy` on the wire) ## `CostRecord` - `id: str` - `tenant_id: uuid.UUID` (`tenantId` on the wire) - `category: str` - `provider_id: str | None` (`providerId` on the wire) (optional) - `run_id: uuid.UUID | None` (`runId` on the wire) (optional) - `agent_id: str | None` (`agentId` on the wire) (optional) - `conversation_id: uuid.UUID | None` (`conversationId` on the wire) (optional) - `quantity: float` - `unit: str` - `cost_usd: float | None` (`costUsd` on the wire) (optional) - `occurred_at: AwareDatetime` (`occurredAt` on the wire) - `metrics: dict[str, Any] | None` (optional) - `attributes: dict[str, Any] | None` (optional) - `call_id: str | None` (`callId` on the wire) (optional) - `project_id: uuid.UUID | None` (`projectId` on the wire) (optional) - `root_run_id: uuid.UUID | None` (`rootRunId` on the wire) (optional) - `parent_run_id: uuid.UUID | None` (`parentRunId` on the wire) (optional) - `agent_version: str | None` (`agentVersion` on the wire) (optional) - `flow_id: str | None` (`flowId` on the wire) (optional) - `node_id: str | None` (`nodeId` on the wire) (optional) - `step: int | None` (optional) - `purpose: str | None` (optional) - `model: str | None` (optional) - `served_model: str | None` (`servedModel` on the wire) (optional) - `fallback: bool | None` (optional) - `status: Optional[Annotated[Union[Literal['ok', 'failed'], str], ]]` (optional) - `usage: kindgi.client.ModelCallTokens | None` (optional) - `thinking: kindgi.client.ModelCallThinking | None` (optional) - `duration_ms: int | None` (`durationMs` on the wire) (optional) - `finish_reason: str | None` (`finishReason` on the wire) (optional) - `provider_request_id: str | None` (`providerRequestId` on the wire) (optional) - `attempts: int | None` (optional) - `error: kindgi.client.ModelCallError | None` (optional) - `raw_usage: kindgi.client.RawUsage | None` (`rawUsage` on the wire) (optional) ## `CostRecordCollectionPage` - `data: list[kindgi.client.CostRecord]` - `next_cursor: str | None` (`nextCursor` on the wire) (optional) - `has_more: bool` (`hasMore` on the wire) ## `CostTokenTotals` - `prompt: int` - `completion: int` - `cache_read: int` (`cacheRead` on the wire) - `cache_write: int` (`cacheWrite` on the wire) - `reasoning: int` ## `CostUsdItem` - `root: float` ## `CreateConnectRequestBody` - `app: str` - `app_id: uuid.UUID | None` (`appId` on the wire) (optional) - `project_id: str | None` (`projectId` on the wire) (optional) - `permissions: list[str]` - `for_: Optional[Literal['me', 'company']]` (`for` on the wire) (optional) ## `CreateConnectionAppBody` Give `oauth2` for an app that signs in with OAuth, `githubApp` for a GitHub App. - `app: str` - `label: str | None` (optional) - `oauth2: kindgi.client.ConnectionOAuth2App | None` (optional) - `github_app: kindgi.client.ConnectionGitHubApp | None` (`githubApp` on the wire) (optional) - `projects: list[str] | None` (optional) ## `CreateCorrectionBody` At least one of `decisions`, `answer` and `items`. `decisions: "wrong"` needs an entry in `shouldCall` or `mustNotCall`; `answer: "wrong"` needs `shouldSay` or `why`. - `run_id: str` (`runId` on the wire) - `decisions: Optional[Literal['right', 'wrong']]` (optional) - `should_call: list[kindgi.client.ShouldCallItem] | None` (`shouldCall` on the wire) (optional) - `must_not_call: list[kindgi.client.MustNotCallItem] | None` (`mustNotCall` on the wire) (optional) - `answer: Optional[Literal['right', 'wrong']]` (optional) - `should_say: str | None` (`shouldSay` on the wire) (optional) - `why: str | None` (optional) - `items: list[kindgi.client.CorrectedAnswerItemInput] | None` (optional) - `corrector_group_id: str | None` (`correctorGroupId` on the wire) (optional) - `participant_id: str | None` (`participantId` on the wire) (optional) ## `CreateCorrectorGroupBody` - `scope: kindgi.client.CorrectorGroupScopeTenant | kindgi.client.CorrectorGroupScopeProject | kindgi.client.CorrectorGroupScopeAgent | kindgi.client.CorrectorGroupScopeUnrecognized` - `name: str` - `weight: float` - `description: str | None` (optional) - `assertable_by: kindgi.client.CorrectorGroupAssertableBy | None` (`assertableBy` on the wire) (optional) ## `CreateMemoryErasureBody` - `root: kindgi.client.MemoryErasureFactSelector | kindgi.client.MemoryErasureSubjectSelector | kindgi.client.MemoryErasureConversationSelector` ## `CreateResourceResult` Shared 201 response body for platform-hierarchy create endpoints — the freshly-assigned id of the created row. - `id: str` ## `CreateServiceAccountBody` - `name: str` - `description: str | None` (optional) - `grants: list[Annotated[kindgi.client.ServiceAccountGrantTenantAdmin | kindgi.client.ServiceAccountGrantTenantMember | kindgi.client.ServiceAccountGrantProjectBody, FieldInfo(annotation=NoneType, required=True, discriminator='kind')]] | None` (optional) ## `CreateSuggestedChangeBody` A suggested change by hand: new settings values (`values`) or a new prompt template (`template`) for one block, in what the project runs (its pin, or Latest). The same change on the same version for the same project is one suggested change (answered with `X-Suggested-Change-Deduped: true`). - `agent_id: str` (`agentId` on the wire) - `project_id: uuid.UUID` (`projectId` on the wire) - `block: str` - `values: dict[str, Any] | None` (optional) - `template: str | None` (optional) - `why: str` - `fixes: list[str] | None` (optional) ## `CreateSystemTenantBody` - `name: str` - `slug: str | None` (optional) - `production_allowed: bool | None` (`productionAllowed` on the wire) (optional) - `test: bool | None` (optional) - `image_prefix: str | None` (`imagePrefix` on the wire) (optional) - `first_admin: kindgi.client.FirstAdmin | None` (`firstAdmin` on the wire) (optional) ## `CreateSystemTenantResult` - `tenant: kindgi.client.SystemTenant` - `first_admin: kindgi.client.FirstAdmin1 | None` (`firstAdmin` on the wire) (optional) - `admin_key: kindgi.client.AdminKey | None` (`adminKey` on the wire) (optional) ## `CreateUserBody` - `display_name: str` (`displayName` on the wire) - `primary_email: str | None` (`primaryEmail` on the wire) (optional) - `send_invite: bool | None` (`sendInvite` on the wire) (optional) ## `CreateWebhookEndpointBody` - `url: AnyUrl` - `events: list[Annotated[Union[Literal['run.finished', 'improve-run.finished', 'approval.requested'], str], ]]` - `filter: kindgi.client.WebhookEndpointFilter | None` (optional) - `secret_ref: kindgi.client.WebhookSecretRef` (`secretRef` on the wire) - `description: str | None` (optional)