@leaflow/sdk 0.139.0 → 0.141.0

package/dist/canopy/v1/generated/validators/currency.zod.js

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/currencyList.zod.d.ts

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/currencyList.zod.js

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/cursorPageRequestResource.zod.d.ts

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/cursorPageRequestResource.zod.js

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/cursorPagination.zod.d.ts

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/cursorPagination.zod.js

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/discount.zod.d.ts

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/discount.zod.js

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/discountList.zod.d.ts

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/discountList.zod.js

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/embeddingRequest.zod.d.ts

@@ -0,0 +1,45 @@
+/**
+ * Generated by orval 🍺
+ * Do not edit manually.
+ * Leaflow Canopy API
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
+ *
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
+ *
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
+ *
+ * ## Authentication
+ *
+ * Every request carries a scoped token; see the `bearerAuth` security scheme.
+ *
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
+ *
+ * ## Conventions
+ *
+ * - **The plaintext of an API key is returned once.** The `secret` appears in the response that issues it and is returned by no other endpoint. A key that has been lost has to be revoked and reissued.
+ * - **Disabling and revoking are distinct.** A disabled key may be enabled again at any time, while a revocation is irreversible and is intended for a key that has been exposed. Neither removes the record, and the usage the key accrued remains readable.
+ * - **A usage query must carry a time range**, spanning no more than 31 days. `from` and `to` carry a timezone offset, which the server does not infer.
+ * - **The request and response bodies do not appear in the ledger.** They are not recorded. Quote the `upstream_request_id` from the response when reporting a problem.
+ * - **Usage may be estimated.** The upstream provider does not report usage on every request. Where it does not, `usage_source` is `estimated` and the figures are derived from the character classes of the payload, to within roughly fifteen per cent.
+ * - **The service is enabled before use.** A project enables the model platform once, through `create-service`; the forwarding endpoints refuse its requests until the service is `active`. Requests are then billed under the service's subscription by the tokens they use, at the rates stated in each model's `pricing`.
+ * OpenAPI spec version: 1.0.0
+ */
+import * as zod from "zod";
+export declare const EmbeddingRequest: zod.ZodObject<{
+    model: zod.ZodString;
+    input: zod.ZodUnion<readonly [zod.ZodString, zod.ZodArray<zod.ZodUnknown>]>;
+    encoding_format: zod.ZodOptional<zod.ZodEnum<{
+        base64: "base64";
+        float: "float";
+    }>>;
+    dimensions: zod.ZodOptional<zod.ZodInt>;
+    user: zod.ZodOptional<zod.ZodString>;
+}, zod.z.core.$strict>;
+export type EmbeddingRequest = zod.input<typeof EmbeddingRequest>;
+export type EmbeddingRequestOutput = zod.output<typeof EmbeddingRequest>;

package/dist/canopy/v1/generated/validators/embeddingRequest.zod.js

@@ -0,0 +1,42 @@
+/**
+ * Generated by orval 🍺
+ * Do not edit manually.
+ * Leaflow Canopy API
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
+ *
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
+ *
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
+ *
+ * ## Authentication
+ *
+ * Every request carries a scoped token; see the `bearerAuth` security scheme.
+ *
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
+ *
+ * ## Conventions
+ *
+ * - **The plaintext of an API key is returned once.** The `secret` appears in the response that issues it and is returned by no other endpoint. A key that has been lost has to be revoked and reissued.
+ * - **Disabling and revoking are distinct.** A disabled key may be enabled again at any time, while a revocation is irreversible and is intended for a key that has been exposed. Neither removes the record, and the usage the key accrued remains readable.
+ * - **A usage query must carry a time range**, spanning no more than 31 days. `from` and `to` carry a timezone offset, which the server does not infer.
+ * - **The request and response bodies do not appear in the ledger.** They are not recorded. Quote the `upstream_request_id` from the response when reporting a problem.
+ * - **Usage may be estimated.** The upstream provider does not report usage on every request. Where it does not, `usage_source` is `estimated` and the figures are derived from the character classes of the payload, to within roughly fifteen per cent.
+ * - **The service is enabled before use.** A project enables the model platform once, through `create-service`; the forwarding endpoints refuse its requests until the service is `active`. Requests are then billed under the service's subscription by the tokens they use, at the rates stated in each model's `pricing`.
+ * OpenAPI spec version: 1.0.0
+ */
+import * as zod from "zod";
+export const EmbeddingRequest = zod.strictObject({
+    model: zod.string().min(1),
+    input: zod
+        .union([zod.string().min(1), zod.array(zod.unknown()).min(1)])
+        .describe("A nonempty string, string array, token integer array, or array of token integer arrays."),
+    encoding_format: zod.enum(["float", "base64"]).optional(),
+    dimensions: zod.int().min(1).optional(),
+    user: zod.string().optional(),
+});

package/dist/canopy/v1/generated/validators/embeddingResponse.zod.d.ts

@@ -0,0 +1,36 @@
+/**
+ * Generated by orval 🍺
+ * Do not edit manually.
+ * Leaflow Canopy API
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
+ *
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
+ *
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
+ *
+ * ## Authentication
+ *
+ * Every request carries a scoped token; see the `bearerAuth` security scheme.
+ *
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
+ *
+ * ## Conventions
+ *
+ * - **The plaintext of an API key is returned once.** The `secret` appears in the response that issues it and is returned by no other endpoint. A key that has been lost has to be revoked and reissued.
+ * - **Disabling and revoking are distinct.** A disabled key may be enabled again at any time, while a revocation is irreversible and is intended for a key that has been exposed. Neither removes the record, and the usage the key accrued remains readable.
+ * - **A usage query must carry a time range**, spanning no more than 31 days. `from` and `to` carry a timezone offset, which the server does not infer.
+ * - **The request and response bodies do not appear in the ledger.** They are not recorded. Quote the `upstream_request_id` from the response when reporting a problem.
+ * - **Usage may be estimated.** The upstream provider does not report usage on every request. Where it does not, `usage_source` is `estimated` and the figures are derived from the character classes of the payload, to within roughly fifteen per cent.
+ * - **The service is enabled before use.** A project enables the model platform once, through `create-service`; the forwarding endpoints refuse its requests until the service is `active`. Requests are then billed under the service's subscription by the tokens they use, at the rates stated in each model's `pricing`.
+ * OpenAPI spec version: 1.0.0
+ */
+import * as zod from "zod";
+export declare const EmbeddingResponse: zod.ZodRecord<zod.ZodString, zod.ZodUnknown>;
+export type EmbeddingResponse = zod.input<typeof EmbeddingResponse>;
+export type EmbeddingResponseOutput = zod.output<typeof EmbeddingResponse>;

package/dist/canopy/v1/generated/validators/embeddingResponse.zod.js

@@ -0,0 +1,34 @@
+/**
+ * Generated by orval 🍺
+ * Do not edit manually.
+ * Leaflow Canopy API
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
+ *
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
+ *
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
+ *
+ * ## Authentication
+ *
+ * Every request carries a scoped token; see the `bearerAuth` security scheme.
+ *
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
+ *
+ * ## Conventions
+ *
+ * - **The plaintext of an API key is returned once.** The `secret` appears in the response that issues it and is returned by no other endpoint. A key that has been lost has to be revoked and reissued.
+ * - **Disabling and revoking are distinct.** A disabled key may be enabled again at any time, while a revocation is irreversible and is intended for a key that has been exposed. Neither removes the record, and the usage the key accrued remains readable.
+ * - **A usage query must carry a time range**, spanning no more than 31 days. `from` and `to` carry a timezone offset, which the server does not infer.
+ * - **The request and response bodies do not appear in the ledger.** They are not recorded. Quote the `upstream_request_id` from the response when reporting a problem.
+ * - **Usage may be estimated.** The upstream provider does not report usage on every request. Where it does not, `usage_source` is `estimated` and the figures are derived from the character classes of the payload, to within roughly fifteen per cent.
+ * - **The service is enabled before use.** A project enables the model platform once, through `create-service`; the forwarding endpoints refuse its requests until the service is `active`. Requests are then billed under the service's subscription by the tokens they use, at the rates stated in each model's `pricing`.
+ * OpenAPI spec version: 1.0.0
+ */
+import * as zod from "zod";
+export const EmbeddingResponse = zod.record(zod.string(), zod.unknown());

package/dist/canopy/v1/generated/validators/entitlement.zod.d.ts

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *

package/dist/canopy/v1/generated/validators/entitlement.zod.js

@@ -2,21 +2,23 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Canopy API
- * Management of API keys, together with the models available and the usage recorded against them.
+ * Canopy model inference, API keys, model catalogue, service enablement and recorded usage.
  *
- * **These endpoints do not perform inference.** Requests are issued to the forwarding address, which accepts the OpenAI, Anthropic and Gemini request formats:
+ * OpenAI-compatible project inference is served on this host at `/v1/chat/completions`,
+ * `/v1/embeddings`, `/v1/models` and `/v1/models/{model}`. Use `https://canopy.leaflow.cloud/v1`
+ * as the OpenAI SDK base URL. Chat completions support `stream=true` SSE.
  *
- * ```
- * POST /v1/chat/completions                    OpenAI
- * POST /v1/messages                            Anthropic
- * POST /v1beta/models/{model}:generateContent  Gemini
- * ```
+ * Management operations use `/api/v1`. The management model catalogue includes pricing;
+ * `/v1/models` returns the OpenAI model format with Canopy capabilities and parameter limits.
  *
  * ## Authentication
  *
  * Every request carries a scoped token; see the `bearerAuth` security scheme.
  *
- * These management endpoints do not accept a Canopy API key. On this userapi host, the OpenAI-compatible `/v1/chat/completions`, `/v1/embeddings`, `/v1/models` and `/v1/models/{model}` endpoints accept the same project token; see the project inference contract. The separate API-key gateway continues to accept Canopy API keys.
+ * Management and project inference accept the same verified IAM scoped token. The token is the
+ * sole billing project authority; `OpenAI-Project` and historical identity headers cannot
+ * override it. These endpoints do not accept a Canopy API key. The separate API-key gateway
+ * accepts Canopy API keys and supports OpenAI, Anthropic and Gemini request formats.
  *
  * ## Conventions
  *