@deeprelay/sdk 0.4.1 → 0.4.2

package/src/apis/RequestsApi.ts

@@ -38,6 +38,7 @@ export interface ExportRequestsRequest {
     clientRequestId?: string;
     from?: Date;
     to?: Date;
+    fallback?: boolean;
 }
 
 export interface GetRequestRequest {
@@ -53,6 +54,7 @@ export interface ListRequestsRequest {
     clientRequestId?: string;
     from?: Date;
     to?: Date;
+    fallback?: boolean;
     cursor?: string;
     limit?: number;
 }
@@ -100,6 +102,10 @@ export class RequestsApi extends runtime.BaseAPI {
             queryParameters['to'] = (requestParameters['to'] as any).toISOString();
         }
 
+        if (requestParameters['fallback'] != null) {
+            queryParameters['fallback'] = requestParameters['fallback'];
+        }
+
         const headerParameters: runtime.HTTPHeaders = {};
 
         if (this.configuration && this.configuration.accessToken) {
@@ -122,7 +128,7 @@ export class RequestsApi extends runtime.BaseAPI {
     }
 
     /**
-     * Streams every request-log row matching the filters as a CSV attachment, oldest columns first: `request_id, client_request_id, occurred_at, surface, model, key_id, key_name, key_prefix, status, http_status, error_code, finish_reason, prompt_tokens, completion_tokens, cached_tokens, reasoning_tokens, quota_units, cost, plan_covered, ttft_ms, duration_ms, ignored_params`. Takes the same filters and visibility rules as `GET /requests`, without the cursor and limit.  The export is counted before anything is written. When the filters match more rows than the export cap, the answer is 422 `export_too_large` and the problem body carries `max_rows`; narrow the time range or add a filter, then export again.  Requires the `serverless:read` scope. Exports have their own, tighter per-key rate limit. 
+     * Streams every request-log row matching the filters as a CSV attachment, oldest columns first: `request_id, client_request_id, occurred_at, surface, model, key_id, key_name, key_prefix, status, http_status, error_code, finish_reason, prompt_tokens, completion_tokens, cached_tokens, reasoning_tokens, quota_units, cost, plan_covered, ttft_ms, duration_ms, ignored_params, requested_model, served_model, fallback_reason`. Takes the same filters and visibility rules as `GET /requests`, without the cursor and limit.  The export is counted before anything is written. When the filters match more rows than the export cap, the answer is 422 `export_too_large` and the problem body carries `max_rows`; narrow the time range or add a filter, then export again.  Requires the `serverless:read` scope. Exports have their own, tighter per-key rate limit. 
      * Export the request log as CSV
      */
     async exportRequestsRaw(requestParameters: ExportRequestsRequest, initOverrides?: RequestInit | runtime.InitOverrideFunction): Promise<runtime.ApiResponse<string>> {
@@ -137,7 +143,7 @@ export class RequestsApi extends runtime.BaseAPI {
     }
 
     /**
-     * Streams every request-log row matching the filters as a CSV attachment, oldest columns first: `request_id, client_request_id, occurred_at, surface, model, key_id, key_name, key_prefix, status, http_status, error_code, finish_reason, prompt_tokens, completion_tokens, cached_tokens, reasoning_tokens, quota_units, cost, plan_covered, ttft_ms, duration_ms, ignored_params`. Takes the same filters and visibility rules as `GET /requests`, without the cursor and limit.  The export is counted before anything is written. When the filters match more rows than the export cap, the answer is 422 `export_too_large` and the problem body carries `max_rows`; narrow the time range or add a filter, then export again.  Requires the `serverless:read` scope. Exports have their own, tighter per-key rate limit. 
+     * Streams every request-log row matching the filters as a CSV attachment, oldest columns first: `request_id, client_request_id, occurred_at, surface, model, key_id, key_name, key_prefix, status, http_status, error_code, finish_reason, prompt_tokens, completion_tokens, cached_tokens, reasoning_tokens, quota_units, cost, plan_covered, ttft_ms, duration_ms, ignored_params, requested_model, served_model, fallback_reason`. Takes the same filters and visibility rules as `GET /requests`, without the cursor and limit.  The export is counted before anything is written. When the filters match more rows than the export cap, the answer is 422 `export_too_large` and the problem body carries `max_rows`; narrow the time range or add a filter, then export again.  Requires the `serverless:read` scope. Exports have their own, tighter per-key rate limit. 
      * Export the request log as CSV
      */
     async exportRequests(requestParameters: ExportRequestsRequest = {}, initOverrides?: RequestInit | runtime.InitOverrideFunction): Promise<string> {
@@ -238,6 +244,10 @@ export class RequestsApi extends runtime.BaseAPI {
             queryParameters['to'] = (requestParameters['to'] as any).toISOString();
         }
 
+        if (requestParameters['fallback'] != null) {
+            queryParameters['fallback'] = requestParameters['fallback'];
+        }
+
         if (requestParameters['cursor'] != null) {
             queryParameters['cursor'] = requestParameters['cursor'];
         }

package/src/models/ApiKey.ts

@@ -20,6 +20,13 @@ import {
     ApiKeyScopesToJSON,
     ApiKeyScopesToJSONTyped,
 } from './ApiKeyScopes';
+import type { ApiKeyFallbackEntry } from './ApiKeyFallbackEntry';
+import {
+    ApiKeyFallbackEntryFromJSON,
+    ApiKeyFallbackEntryFromJSONTyped,
+    ApiKeyFallbackEntryToJSON,
+    ApiKeyFallbackEntryToJSONTyped,
+} from './ApiKeyFallbackEntry';
 
 /**
  * An API key's metadata and limits. Never carries the secret.
@@ -87,6 +94,24 @@ export interface ApiKey {
      * @memberof ApiKey
      */
     staleModels?: Array<string>;
+    /**
+     * Backup models in priority order, applied to every model this key calls. Null means model fallback is off (the default). If every provider for the requested model fails before the first token with a temporary error, the next entry is tried. Entries are kept when a model leaves the plan roster or the catalog and are skipped at request time; see `fallback_entries`. Not applied to Liquid.
+     * @type {Array<string>}
+     * @memberof ApiKey
+     */
+    fallbackModels: Array<string> | null;
+    /**
+     * Present when `fallback_models` is set: one entry per backup, in order, explaining it as of now.
+     * @type {Array<ApiKeyFallbackEntry>}
+     * @memberof ApiKey
+     */
+    fallbackEntries?: Array<ApiKeyFallbackEntry>;
+    /**
+     * The model this key requested most over the last 30 days, which `above_primary` is measured against. Present only for a key with `fallback_models` and recent usage.
+     * @type {string}
+     * @memberof ApiKey
+     */
+    primaryModel?: string;
     /**
      * 
      * @type {number}
@@ -136,6 +161,7 @@ export function instanceOfApiKey(value: object): value is ApiKey {
     if (!('scopes' in value) || value['scopes'] === undefined) return false;
     if ((!('budgetCents' in (value as Record<string, any>)) && !('budget_cents' in (value as Record<string, any>))) || ((value as Record<string, any>)['budgetCents'] === undefined && (value as Record<string, any>)['budget_cents'] === undefined)) return false;
     if ((!('allowedModels' in (value as Record<string, any>)) && !('allowed_models' in (value as Record<string, any>))) || ((value as Record<string, any>)['allowedModels'] === undefined && (value as Record<string, any>)['allowed_models'] === undefined)) return false;
+    if ((!('fallbackModels' in (value as Record<string, any>)) && !('fallback_models' in (value as Record<string, any>))) || ((value as Record<string, any>)['fallbackModels'] === undefined && (value as Record<string, any>)['fallback_models'] === undefined)) return false;
     if ((!('rateLimitRps' in (value as Record<string, any>)) && !('rate_limit_rps' in (value as Record<string, any>))) || ((value as Record<string, any>)['rateLimitRps'] === undefined && (value as Record<string, any>)['rate_limit_rps'] === undefined)) return false;
     if ((!('rateLimitBurst' in (value as Record<string, any>)) && !('rate_limit_burst' in (value as Record<string, any>))) || ((value as Record<string, any>)['rateLimitBurst'] === undefined && (value as Record<string, any>)['rate_limit_burst'] === undefined)) return false;
     if ((!('expiresAt' in (value as Record<string, any>)) && !('expires_at' in (value as Record<string, any>))) || ((value as Record<string, any>)['expiresAt'] === undefined && (value as Record<string, any>)['expires_at'] === undefined)) return false;
@@ -164,6 +190,9 @@ export function ApiKeyFromJSONTyped(json: any, ignoreDiscriminator: boolean): Ap
         'budgetResetsAt': json['budget_resets_at'] == null ? undefined : (new Date(json['budget_resets_at'])),
         'allowedModels': json['allowed_models'] == null ? null : json['allowed_models'],
         'staleModels': json['stale_models'] == null ? undefined : json['stale_models'],
+        'fallbackModels': json['fallback_models'] == null ? null : json['fallback_models'],
+        'fallbackEntries': json['fallback_entries'] == null ? undefined : ((json['fallback_entries'] as Array<any>).map(ApiKeyFallbackEntryFromJSON)),
+        'primaryModel': json['primary_model'] == null ? undefined : json['primary_model'],
         'rateLimitRps': json['rate_limit_rps'],
         'rateLimitBurst': json['rate_limit_burst'],
         'expiresAt': (json['expires_at'] == null ? null : new Date(json['expires_at'])),
@@ -194,6 +223,9 @@ export function ApiKeyToJSONTyped(value?: ApiKey | null, ignoreDiscriminator: bo
         'budget_resets_at': value['budgetResetsAt'] == null ? value['budgetResetsAt'] : value['budgetResetsAt'].toISOString(),
         'allowed_models': value['allowedModels'],
         'stale_models': value['staleModels'],
+        'fallback_models': value['fallbackModels'],
+        'fallback_entries': value['fallbackEntries'] == null ? undefined : ((value['fallbackEntries'] as Array<any>).map(ApiKeyFallbackEntryToJSON)),
+        'primary_model': value['primaryModel'],
         'rate_limit_rps': value['rateLimitRps'],
         'rate_limit_burst': value['rateLimitBurst'],
         'expires_at': value['expiresAt'] == null ? value['expiresAt'] : value['expiresAt'].toISOString(),

package/src/models/ApiKeyCreate.ts

@@ -43,6 +43,12 @@ export interface ApiKeyCreate {
      * @memberof ApiKeyCreate
      */
     allowedModels?: Array<string>;
+    /**
+     * Backup models in priority order. Omit it to leave model fallback off. Each entry must be a chat model in the catalog and, when the key has `allowed_models`, on that list. Duplicates are removed. An empty array, more than 3 models or an invalid entry is a 422 `validation_failed` naming the `fallback_models` field.
+     * @type {Array<string>}
+     * @memberof ApiKeyCreate
+     */
+    fallbackModels?: Array<string>;
     /**
      * 
      * @type {number}
@@ -86,6 +92,7 @@ export function ApiKeyCreateFromJSONTyped(json: any, ignoreDiscriminator: boolea
         'scopes': json['scopes'],
         'budgetCents': json['budget_cents'] == null ? undefined : json['budget_cents'],
         'allowedModels': json['allowed_models'] == null ? undefined : json['allowed_models'],
+        'fallbackModels': json['fallback_models'] == null ? undefined : json['fallback_models'],
         'rateLimitRps': json['rate_limit_rps'] == null ? undefined : json['rate_limit_rps'],
         'rateLimitBurst': json['rate_limit_burst'] == null ? undefined : json['rate_limit_burst'],
         'expiresAt': json['expires_at'] == null ? undefined : (new Date(json['expires_at'])),
@@ -107,6 +114,7 @@ export function ApiKeyCreateToJSONTyped(value?: ApiKeyCreate | null, ignoreDiscr
         'scopes': value['scopes'],
         'budget_cents': value['budgetCents'],
         'allowed_models': value['allowedModels'],
+        'fallback_models': value['fallbackModels'],
         'rate_limit_rps': value['rateLimitRps'],
         'rate_limit_burst': value['rateLimitBurst'],
         'expires_at': value['expiresAt'] == null ? value['expiresAt'] : value['expiresAt'].toISOString(),

package/src/models/ApiKeyCreated.ts

@@ -20,6 +20,13 @@ import {
     ApiKeyScopesToJSON,
     ApiKeyScopesToJSONTyped,
 } from './ApiKeyScopes';
+import type { ApiKeyFallbackEntry } from './ApiKeyFallbackEntry';
+import {
+    ApiKeyFallbackEntryFromJSON,
+    ApiKeyFallbackEntryFromJSONTyped,
+    ApiKeyFallbackEntryToJSON,
+    ApiKeyFallbackEntryToJSONTyped,
+} from './ApiKeyFallbackEntry';
 
 /**
  * The new key's metadata plus its secret (`key`), returned exactly once. The same fields as `ApiKey`, plus `key`.
@@ -87,6 +94,24 @@ export interface ApiKeyCreated {
      * @memberof ApiKeyCreated
      */
     staleModels?: Array<string>;
+    /**
+     * Backup models in priority order, applied to every model this key calls. Null means model fallback is off (the default). If every provider for the requested model fails before the first token with a temporary error, the next entry is tried. Entries are kept when a model leaves the plan roster or the catalog and are skipped at request time; see `fallback_entries`. Not applied to Liquid.
+     * @type {Array<string>}
+     * @memberof ApiKeyCreated
+     */
+    fallbackModels: Array<string> | null;
+    /**
+     * Present when `fallback_models` is set: one entry per backup, in order, explaining it as of now.
+     * @type {Array<ApiKeyFallbackEntry>}
+     * @memberof ApiKeyCreated
+     */
+    fallbackEntries?: Array<ApiKeyFallbackEntry>;
+    /**
+     * The model this key requested most over the last 30 days, which `above_primary` is measured against. Present only for a key with `fallback_models` and recent usage.
+     * @type {string}
+     * @memberof ApiKeyCreated
+     */
+    primaryModel?: string;
     /**
      * 
      * @type {number}
@@ -142,6 +167,7 @@ export function instanceOfApiKeyCreated(value: object): value is ApiKeyCreated {
     if (!('scopes' in value) || value['scopes'] === undefined) return false;
     if ((!('budgetCents' in (value as Record<string, any>)) && !('budget_cents' in (value as Record<string, any>))) || ((value as Record<string, any>)['budgetCents'] === undefined && (value as Record<string, any>)['budget_cents'] === undefined)) return false;
     if ((!('allowedModels' in (value as Record<string, any>)) && !('allowed_models' in (value as Record<string, any>))) || ((value as Record<string, any>)['allowedModels'] === undefined && (value as Record<string, any>)['allowed_models'] === undefined)) return false;
+    if ((!('fallbackModels' in (value as Record<string, any>)) && !('fallback_models' in (value as Record<string, any>))) || ((value as Record<string, any>)['fallbackModels'] === undefined && (value as Record<string, any>)['fallback_models'] === undefined)) return false;
     if ((!('rateLimitRps' in (value as Record<string, any>)) && !('rate_limit_rps' in (value as Record<string, any>))) || ((value as Record<string, any>)['rateLimitRps'] === undefined && (value as Record<string, any>)['rate_limit_rps'] === undefined)) return false;
     if ((!('rateLimitBurst' in (value as Record<string, any>)) && !('rate_limit_burst' in (value as Record<string, any>))) || ((value as Record<string, any>)['rateLimitBurst'] === undefined && (value as Record<string, any>)['rate_limit_burst'] === undefined)) return false;
     if ((!('expiresAt' in (value as Record<string, any>)) && !('expires_at' in (value as Record<string, any>))) || ((value as Record<string, any>)['expiresAt'] === undefined && (value as Record<string, any>)['expires_at'] === undefined)) return false;
@@ -171,6 +197,9 @@ export function ApiKeyCreatedFromJSONTyped(json: any, ignoreDiscriminator: boole
         'budgetResetsAt': json['budget_resets_at'] == null ? undefined : (new Date(json['budget_resets_at'])),
         'allowedModels': json['allowed_models'] == null ? null : json['allowed_models'],
         'staleModels': json['stale_models'] == null ? undefined : json['stale_models'],
+        'fallbackModels': json['fallback_models'] == null ? null : json['fallback_models'],
+        'fallbackEntries': json['fallback_entries'] == null ? undefined : ((json['fallback_entries'] as Array<any>).map(ApiKeyFallbackEntryFromJSON)),
+        'primaryModel': json['primary_model'] == null ? undefined : json['primary_model'],
         'rateLimitRps': json['rate_limit_rps'],
         'rateLimitBurst': json['rate_limit_burst'],
         'expiresAt': (json['expires_at'] == null ? null : new Date(json['expires_at'])),
@@ -202,6 +231,9 @@ export function ApiKeyCreatedToJSONTyped(value?: ApiKeyCreated | null, ignoreDis
         'budget_resets_at': value['budgetResetsAt'] == null ? value['budgetResetsAt'] : value['budgetResetsAt'].toISOString(),
         'allowed_models': value['allowedModels'],
         'stale_models': value['staleModels'],
+        'fallback_models': value['fallbackModels'],
+        'fallback_entries': value['fallbackEntries'] == null ? undefined : ((value['fallbackEntries'] as Array<any>).map(ApiKeyFallbackEntryToJSON)),
+        'primary_model': value['primaryModel'],
         'rate_limit_rps': value['rateLimitRps'],
         'rate_limit_burst': value['rateLimitBurst'],
         'expires_at': value['expiresAt'] == null ? value['expiresAt'] : value['expiresAt'].toISOString(),

package/src/models/ApiKeyFallbackEntry.ts

@@ -0,0 +1,114 @@
+/* tslint:disable */
+/* eslint-disable */
+/**
+ * deeprelay Public Developer API
+ * Programmatic API for deeprelay. Authenticate with API keys minted from the dashboard. See https://deeprelay.ai/docs for details.  Keys minted from the dashboard or `deeprelay login` belong to the Production server, https://api.deeprelay.ai/v1. The Demo server is an internal staging environment with its own key store.  Versioning: `/v1` is the API version. `info.version` is the revision of this `/v1` document, not a separate API.  OpenAI compatibility: chat completions follow the OpenAI request and response shapes, but not every OpenAI parameter is carried. Some are rejected with 400 `unsupported_parameter` and any parameter not declared in this spec is accepted and ignored. The full table is at https://deeprelay.ai/docs/api/inference#openai-compatibility. 
+ *
+ * The version of the OpenAPI document: 1.0.0
+ * Contact: support@deeprelay.ai
+ *
+ * NOTE: This class is auto generated by OpenAPI Generator (https://openapi-generator.tech).
+ * https://openapi-generator.tech
+ * Do not edit the class manually.
+ */
+
+import { mapValues } from '../runtime';
+/**
+ * One backup model on a key, as it stands now.
+ * @export
+ * @interface ApiKeyFallbackEntry
+ */
+export interface ApiKeyFallbackEntry {
+    /**
+     * 
+     * @type {string}
+     * @memberof ApiKeyFallbackEntry
+     */
+    model: string;
+    /**
+     * `ok`: usable as a backup. `wallet_only`: the model is not on the subscription plan roster, so it backs up only requests billed to the wallet; a plan-covered request skips it. `unavailable`: the model is no longer in the catalog and is always skipped.
+     * @type {ApiKeyFallbackEntryStateEnum}
+     * @memberof ApiKeyFallbackEntry
+     */
+    state: ApiKeyFallbackEntryStateEnum;
+    /**
+     * Current input price per 1M tokens, in micro-cents. 0 when unknown or unavailable.
+     * @type {number}
+     * @memberof ApiKeyFallbackEntry
+     */
+    inputPer1mTokensMicrocents: number;
+    /**
+     * Current output price per 1M tokens, in micro-cents. 0 when unknown or unavailable.
+     * @type {number}
+     * @memberof ApiKeyFallbackEntry
+     */
+    outputPer1mTokensMicrocents: number;
+    /**
+     * True when this backup's input plus output price is above `primary_model`'s. Pricier backups are allowed; a request served by one is billed at its own price, and the key budget is checked at that price.
+     * @type {boolean}
+     * @memberof ApiKeyFallbackEntry
+     */
+    abovePrimary: boolean;
+}
+
+
+/**
+ * @export
+ */
+export const ApiKeyFallbackEntryStateEnum = {
+    Ok: 'ok',
+    WalletOnly: 'wallet_only',
+    Unavailable: 'unavailable'
+} as const;
+export type ApiKeyFallbackEntryStateEnum = typeof ApiKeyFallbackEntryStateEnum[keyof typeof ApiKeyFallbackEntryStateEnum];
+
+
+/**
+ * Check if a given object implements the ApiKeyFallbackEntry interface.
+ */
+export function instanceOfApiKeyFallbackEntry(value: object): value is ApiKeyFallbackEntry {
+    if (!('model' in value) || value['model'] === undefined) return false;
+    if (!('state' in value) || value['state'] === undefined) return false;
+    if ((!('inputPer1mTokensMicrocents' in (value as Record<string, any>)) && !('input_per_1m_tokens_microcents' in (value as Record<string, any>))) || ((value as Record<string, any>)['inputPer1mTokensMicrocents'] === undefined && (value as Record<string, any>)['input_per_1m_tokens_microcents'] === undefined)) return false;
+    if ((!('outputPer1mTokensMicrocents' in (value as Record<string, any>)) && !('output_per_1m_tokens_microcents' in (value as Record<string, any>))) || ((value as Record<string, any>)['outputPer1mTokensMicrocents'] === undefined && (value as Record<string, any>)['output_per_1m_tokens_microcents'] === undefined)) return false;
+    if ((!('abovePrimary' in (value as Record<string, any>)) && !('above_primary' in (value as Record<string, any>))) || ((value as Record<string, any>)['abovePrimary'] === undefined && (value as Record<string, any>)['above_primary'] === undefined)) return false;
+    return true;
+}
+
+export function ApiKeyFallbackEntryFromJSON(json: any): ApiKeyFallbackEntry {
+    return ApiKeyFallbackEntryFromJSONTyped(json, false);
+}
+
+export function ApiKeyFallbackEntryFromJSONTyped(json: any, ignoreDiscriminator: boolean): ApiKeyFallbackEntry {
+    if (json == null) {
+        return json;
+    }
+    return {
+        
+        'model': json['model'],
+        'state': json['state'],
+        'inputPer1mTokensMicrocents': json['input_per_1m_tokens_microcents'],
+        'outputPer1mTokensMicrocents': json['output_per_1m_tokens_microcents'],
+        'abovePrimary': json['above_primary'],
+    };
+}
+
+export function ApiKeyFallbackEntryToJSON(json: any): ApiKeyFallbackEntry {
+    return ApiKeyFallbackEntryToJSONTyped(json, false);
+}
+
+export function ApiKeyFallbackEntryToJSONTyped(value?: ApiKeyFallbackEntry | null, ignoreDiscriminator: boolean = false): any {
+    if (value == null) {
+        return value;
+    }
+
+    return {
+        
+        'model': value['model'],
+        'state': value['state'],
+        'input_per_1m_tokens_microcents': value['inputPer1mTokensMicrocents'],
+        'output_per_1m_tokens_microcents': value['outputPer1mTokensMicrocents'],
+        'above_primary': value['abovePrimary'],
+    };
+}
+

package/src/models/ApiKeyUpdate.ts

@@ -14,7 +14,7 @@
 
 import { mapValues } from '../runtime';
 /**
- * Send only the fields to change. `null` clears `budget_cents`, `allowed_models` or `expires_at`. `scopes` cannot be changed.
+ * Send only the fields to change. `null` clears `budget_cents`, `allowed_models`, `fallback_models` or `expires_at`. `scopes` cannot be changed.
  * @export
  * @interface ApiKeyUpdate
  */
@@ -37,6 +37,12 @@ export interface ApiKeyUpdate {
      * @memberof ApiKeyUpdate
      */
     allowedModels?: Array<string> | null;
+    /**
+     * Replace the backup list (priority order), or `null` to turn model fallback off. Same rules as on create, judged against the key's allowlist after this update. An empty array is a 422 `validation_failed`; send `null` instead.
+     * @type {Array<string>}
+     * @memberof ApiKeyUpdate
+     */
+    fallbackModels?: Array<string> | null;
     /**
      * 
      * @type {number}
@@ -77,6 +83,7 @@ export function ApiKeyUpdateFromJSONTyped(json: any, ignoreDiscriminator: boolea
         'name': json['name'] == null ? undefined : json['name'],
         'budgetCents': json['budget_cents'] === undefined ? undefined : json['budget_cents'] === null ? null : json['budget_cents'],
         'allowedModels': json['allowed_models'] === undefined ? undefined : json['allowed_models'] === null ? null : json['allowed_models'],
+        'fallbackModels': json['fallback_models'] === undefined ? undefined : json['fallback_models'] === null ? null : json['fallback_models'],
         'rateLimitRps': json['rate_limit_rps'] == null ? undefined : json['rate_limit_rps'],
         'rateLimitBurst': json['rate_limit_burst'] == null ? undefined : json['rate_limit_burst'],
         'expiresAt': json['expires_at'] === undefined ? undefined : json['expires_at'] === null ? null : (new Date(json['expires_at'])),
@@ -97,6 +104,7 @@ export function ApiKeyUpdateToJSONTyped(value?: ApiKeyUpdate | null, ignoreDiscr
         'name': value['name'],
         'budget_cents': value['budgetCents'],
         'allowed_models': value['allowedModels'],
+        'fallback_models': value['fallbackModels'],
         'rate_limit_rps': value['rateLimitRps'],
         'rate_limit_burst': value['rateLimitBurst'],
         'expires_at': value['expiresAt'] == null ? value['expiresAt'] : value['expiresAt'].toISOString(),

package/src/models/ChatCompletionRequest.ts

@@ -61,6 +61,12 @@ export interface ChatCompletionRequest {
      * @memberof ChatCompletionRequest
      */
     model: string;
+    /**
+     * OpenRouter-compatible backup models for this request. `model` is the requested model and `models` lists the backups in priority order; when `model` is empty, the first entry of `models` is the requested model. An entry naming the requested model is ignored. At most 3 backups. The list replaces the API key's `fallback_models` for this request, and an empty array turns model fallback off for this request. Each backup must still pass the key's model allowlist, the request's billing mode and its capability needs, or it is skipped. Sending both `models` and the X-Deeprelay-Fallback-Models header is a 400 `invalid_request`. Chat completions only; on `/messages` and `/responses` use the header.
+     * @type {Array<string>}
+     * @memberof ChatCompletionRequest
+     */
+    models?: Array<string> | null;
     /**
      * 
      * @type {Array<ChatMessage>}
@@ -186,6 +192,7 @@ export function ChatCompletionRequestFromJSONTyped(json: any, ignoreDiscriminato
     return {
         
         'model': json['model'],
+        'models': json['models'] === undefined ? undefined : json['models'] === null ? null : json['models'],
         'messages': ((json['messages'] as Array<any>).map(ChatMessageFromJSON)),
         'temperature': json['temperature'] == null ? undefined : json['temperature'],
         'maxTokens': json['max_tokens'] == null ? undefined : json['max_tokens'],
@@ -217,6 +224,7 @@ export function ChatCompletionRequestToJSONTyped(value?: ChatCompletionRequest |
     return {
         
         'model': value['model'],
+        'models': value['models'],
         'messages': ((value['messages'] as Array<any>).map(ChatMessageToJSON)),
         'temperature': value['temperature'],
         'max_tokens': value['maxTokens'],

package/src/models/CreateWebhookEndpointRequest.ts

@@ -47,7 +47,8 @@ export const CreateWebhookEndpointRequestEventTypesEnum = {
     BalanceLow: 'balance.low',
     SubscriptionQuotaThreshold: 'subscription_quota.threshold',
     ApiKeyBudgetThreshold: 'api_key_budget.threshold',
-    MemberSpendThreshold: 'member_spend.threshold'
+    MemberSpendThreshold: 'member_spend.threshold',
+    PlanRosterModelRemoved: 'plan_roster.model_removed'
 } as const;
 export type CreateWebhookEndpointRequestEventTypesEnum = typeof CreateWebhookEndpointRequestEventTypesEnum[keyof typeof CreateWebhookEndpointRequestEventTypesEnum];
 

package/src/models/RequestAttempt.ts

@@ -0,0 +1,121 @@
+/* tslint:disable */
+/* eslint-disable */
+/**
+ * deeprelay Public Developer API
+ * Programmatic API for deeprelay. Authenticate with API keys minted from the dashboard. See https://deeprelay.ai/docs for details.  Keys minted from the dashboard or `deeprelay login` belong to the Production server, https://api.deeprelay.ai/v1. The Demo server is an internal staging environment with its own key store.  Versioning: `/v1` is the API version. `info.version` is the revision of this `/v1` document, not a separate API.  OpenAI compatibility: chat completions follow the OpenAI request and response shapes, but not every OpenAI parameter is carried. Some are rejected with 400 `unsupported_parameter` and any parameter not declared in this spec is accepted and ignored. The full table is at https://deeprelay.ai/docs/api/inference#openai-compatibility. 
+ *
+ * The version of the OpenAPI document: 1.0.0
+ * Contact: support@deeprelay.ai
+ *
+ * NOTE: This class is auto generated by OpenAPI Generator (https://openapi-generator.tech).
+ * https://openapi-generator.tech
+ * Do not edit the class manually.
+ */
+
+import { mapValues } from '../runtime';
+/**
+ * One model in a request's backup-model trail.
+ * @export
+ * @interface RequestAttempt
+ */
+export interface RequestAttempt {
+    /**
+     * 
+     * @type {string}
+     * @memberof RequestAttempt
+     */
+    model: string;
+    /**
+     * `ok`: this model served. `failed`: it was dialled and every provider failed before the first token. `skipped`: it was passed over without being dialled.
+     * @type {RequestAttemptOutcomeEnum}
+     * @memberof RequestAttempt
+     */
+    outcome: RequestAttemptOutcomeEnum;
+    /**
+     * Why the attempt failed (a public error code such as `upstream_timeout`) or was skipped: `model_not_allowed_for_key`, `subscription_quota_exhausted`, `api_key_budget_exceeded`, `member_spend_limit_reached`, `insufficient_balance`, `daily_limit_reached`, `spending_limit_reached`, `wallet_only` (not on the plan roster, for a plan-covered request), `unavailable`, `capability_mismatch`, `context_window_too_small`, `breaker_open` or `deadline_exhausted` (too little time was left to start another model).
+     * @type {string}
+     * @memberof RequestAttempt
+     */
+    reason?: string;
+    /**
+     * How long this model's attempt took, in milliseconds.
+     * @type {number}
+     * @memberof RequestAttempt
+     */
+    durationMs: number;
+    /**
+     * Time to first token, for the model that served a streamed call.
+     * @type {number}
+     * @memberof RequestAttempt
+     */
+    ttftMs?: number;
+    /**
+     * How many upstream attempts were made for this model. 0 for a skipped model.
+     * @type {number}
+     * @memberof RequestAttempt
+     */
+    tries: number;
+}
+
+
+/**
+ * @export
+ */
+export const RequestAttemptOutcomeEnum = {
+    Ok: 'ok',
+    Failed: 'failed',
+    Skipped: 'skipped'
+} as const;
+export type RequestAttemptOutcomeEnum = typeof RequestAttemptOutcomeEnum[keyof typeof RequestAttemptOutcomeEnum];
+
+
+/**
+ * Check if a given object implements the RequestAttempt interface.
+ */
+export function instanceOfRequestAttempt(value: object): value is RequestAttempt {
+    if (!('model' in value) || value['model'] === undefined) return false;
+    if (!('outcome' in value) || value['outcome'] === undefined) return false;
+    if ((!('durationMs' in (value as Record<string, any>)) && !('duration_ms' in (value as Record<string, any>))) || ((value as Record<string, any>)['durationMs'] === undefined && (value as Record<string, any>)['duration_ms'] === undefined)) return false;
+    if (!('tries' in value) || value['tries'] === undefined) return false;
+    return true;
+}
+
+export function RequestAttemptFromJSON(json: any): RequestAttempt {
+    return RequestAttemptFromJSONTyped(json, false);
+}
+
+export function RequestAttemptFromJSONTyped(json: any, ignoreDiscriminator: boolean): RequestAttempt {
+    if (json == null) {
+        return json;
+    }
+    return {
+        
+        'model': json['model'],
+        'outcome': json['outcome'],
+        'reason': json['reason'] == null ? undefined : json['reason'],
+        'durationMs': json['duration_ms'],
+        'ttftMs': json['ttft_ms'] == null ? undefined : json['ttft_ms'],
+        'tries': json['tries'],
+    };
+}
+
+export function RequestAttemptToJSON(json: any): RequestAttempt {
+    return RequestAttemptToJSONTyped(json, false);
+}
+
+export function RequestAttemptToJSONTyped(value?: RequestAttempt | null, ignoreDiscriminator: boolean = false): any {
+    if (value == null) {
+        return value;
+    }
+
+    return {
+        
+        'model': value['model'],
+        'outcome': value['outcome'],
+        'reason': value['reason'],
+        'duration_ms': value['durationMs'],
+        'ttft_ms': value['ttftMs'],
+        'tries': value['tries'],
+    };
+}
+

package/src/models/RequestLogEntry.ts

@@ -13,6 +13,14 @@
  */
 
 import { mapValues } from '../runtime';
+import type { RequestAttempt } from './RequestAttempt';
+import {
+    RequestAttemptFromJSON,
+    RequestAttemptFromJSONTyped,
+    RequestAttemptToJSON,
+    RequestAttemptToJSONTyped,
+} from './RequestAttempt';
+
 /**
  * One inference call as the request log records it. Never carries prompt or completion content. Optional numeric fields are omitted when they do not apply (for example token counts on a rejected call).
  * @export
@@ -181,6 +189,30 @@ export interface RequestLogEntry {
      * @memberof RequestLogEntry
      */
     suppressedCount?: number;
+    /**
+     * The model the request asked for. Always present; equal to `model` unless a backup model was tried.
+     * @type {string}
+     * @memberof RequestLogEntry
+     */
+    requestedModel?: string;
+    /**
+     * The model that produced the response. Omitted when no model served (a failed or rejected call). When a backup served, `model` is this model and the row is billed at its price.
+     * @type {string}
+     * @memberof RequestLogEntry
+     */
+    servedModel?: string;
+    /**
+     * Set when backup models were tried: the requested model's public error code (for example `upstream_timeout`). Rows with it set are the ones `fallback=true` returns.
+     * @type {string}
+     * @memberof RequestLogEntry
+     */
+    fallbackReason?: string;
+    /**
+     * Present when backup models were tried: every model the request went through, in order, including skipped backups. Upstream providers are never named.
+     * @type {Array<RequestAttempt>}
+     * @memberof RequestLogEntry
+     */
+    attempts?: Array<RequestAttempt>;
 }
 
 
@@ -261,6 +293,10 @@ export function RequestLogEntryFromJSONTyped(json: any, ignoreDiscriminator: boo
         'imageCount': json['image_count'] == null ? undefined : json['image_count'],
         'videoSeconds': json['video_seconds'] == null ? undefined : json['video_seconds'],
         'suppressedCount': json['suppressed_count'] == null ? undefined : json['suppressed_count'],
+        'requestedModel': json['requested_model'] == null ? undefined : json['requested_model'],
+        'servedModel': json['served_model'] == null ? undefined : json['served_model'],
+        'fallbackReason': json['fallback_reason'] == null ? undefined : json['fallback_reason'],
+        'attempts': json['attempts'] == null ? undefined : ((json['attempts'] as Array<any>).map(RequestAttemptFromJSON)),
     };
 }
 
@@ -302,6 +338,10 @@ export function RequestLogEntryToJSONTyped(value?: RequestLogEntry | null, ignor
         'image_count': value['imageCount'],
         'video_seconds': value['videoSeconds'],
         'suppressed_count': value['suppressedCount'],
+        'requested_model': value['requestedModel'],
+        'served_model': value['servedModel'],
+        'fallback_reason': value['fallbackReason'],
+        'attempts': value['attempts'] == null ? undefined : ((value['attempts'] as Array<any>).map(RequestAttemptToJSON)),
     };
 }
 

package/src/models/WebhookEndpoint.ts

@@ -71,7 +71,8 @@ export const WebhookEndpointEventTypesEnum = {
     BalanceLow: 'balance.low',
     SubscriptionQuotaThreshold: 'subscription_quota.threshold',
     ApiKeyBudgetThreshold: 'api_key_budget.threshold',
-    MemberSpendThreshold: 'member_spend.threshold'
+    MemberSpendThreshold: 'member_spend.threshold',
+    PlanRosterModelRemoved: 'plan_roster.model_removed'
 } as const;
 export type WebhookEndpointEventTypesEnum = typeof WebhookEndpointEventTypesEnum[keyof typeof WebhookEndpointEventTypesEnum];
 

package/src/models/index.ts

@@ -6,6 +6,7 @@ export * from './AlertSettingsUpdate';
 export * from './ApiKey';
 export * from './ApiKeyCreate';
 export * from './ApiKeyCreated';
+export * from './ApiKeyFallbackEntry';
 export * from './ApiKeyList';
 export * from './ApiKeyScopes';
 export * from './ApiKeyUpdate';
@@ -45,6 +46,7 @@ export * from './PeakWindow';
 export * from './Problem';
 export * from './RefereeStatus';
 export * from './Referral';
+export * from './RequestAttempt';
 export * from './RequestLogEntry';
 export * from './RequestLogPage';
 export * from './ResponseFormatJSONObject';

package/src/runtime.ts

@@ -75,7 +75,7 @@ export class Configuration {
     }
 
     get headers(): HTTPHeaders | undefined {
-        return { 'User-Agent': 'deeprelay-sdk-ts/0.4.1', ...this.configuration.headers };
+        return { 'User-Agent': 'deeprelay-sdk-ts/0.4.2', ...this.configuration.headers };
     }
 
     get credentials(): RequestCredentials | undefined {