@leaflow/sdk 0.0.0-dev.358.gbb34078 → 0.0.0-dev.361.gefdc428

package/dist/assistant/v1/generated/validators/clientActionRequest.zod.d.ts

@@ -2,65 +2,41 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Assistant API
- * The AI assistant inside the console: conversations, messages, tool calls and their approval, attachments, and connecting the assistant to instant messaging platforms such as WeChat, Feishu and Telegram.
+ * Conversations with agents that operate the Leaflow platform, plus their folders, tasks, files and
+ * chat-app channels.
  *
- * ## A conversation
+ * A conversation is a thread. A person talks to the main thread; the agent in it can start subagents,
+ * which are threads too and run on their own: stopping the main thread does not stop them, and each
+ * subagent tells its parent when it ends. A thread runs in turns; each turn records the model and
+ * reasoning effort it used and its token usage.
  *
- * The path one conversation takes, end to end. It names the operations that path needs, not all of them — the full list is below, and it also covers folders, attachments, memories, skills and the instant messaging side.
+ * Models come from the Canopy catalogue. A new thread uses the model it names, the caller's default
+ * from /api/v1/settings, or the first catalogue model that can run an agent. Reasoning models take one
+ * of three efforts — low, medium or high — which are sent to the model unchanged.
  *
- * 1. `POST /api/v1/threads` creates one;
- * 2. `POST /api/v1/threads/{thread}/messages` sends a message and returns a `turnId` at once; the assistant runs in the background;
- * 3. `GET /api/v1/threads/{thread}` returns the conversation document to render, whose `stream` is where live output arrives;
- * 4. When the assistant needs the user, the document's `wait` says which kind: approving a batch of tool calls (`POST /api/v1/threads/{thread}/approvals/{batch}`), or answering a few questions (`POST /api/v1/threads/{thread}/questions/{item}`). It continues on its own once that is handled.
+ * Read a thread's state with get-thread and its transcript with list-thread-items. Follow both live
+ * over the WebSocket described under /api/v1/realtime/tickets. Slash commands such as /status and
+ * /compact are listed by list-commands and work the same in the console and in chat apps.
  *
- * Sending a file takes two steps: `POST /api/v1/attachments` for an id, then put that id in `attachmentIds` when sending the message. Images, text and arbitrary binaries all go through the same two steps; what differs is how the assistant reads them afterwards, which the upload response's `kind` says.
- *
- * ## Models
- *
- * Canopy owns the model catalogue and capabilities. Model ids in this API, conversation preferences and historical entries are Canopy model ids, with no separate assistant model registration. GET /api/v1/models returns the allowed, usable choices and allowModelSelection. Hide the selector when selection is disabled; the service enforces the same rule.
- *
- * preferredModelId is the user's choice. Null follows the platform default, including later changes. Disabled selection preserves but ignores preferences for new turns. A model removed from the allowlist, retired in Canopy or otherwise unusable falls back to the default without clearing the preference.
- *
- * Users can switch an existing conversation through PATCH /api/v1/threads/{thread} with preferredModelId. This preserves the transcript, starts no turn and makes no inference request. Null returns to following the default. Changes apply from the next turn; a running or waiting turn and messages queued into it retain their captured choices and fallback policy.
- *
- * If the chosen model cannot serve a request, the service may safely fall back to the default without confirmation or an additional switching message. If the default also fails, the turn fails normally; no third model is selected for ordinary answers. Fallback preserves executed tool results and never splices answers from different models. Invalid input, cancellation and a declined tool call do not trigger model fallback. Each entry's model records its actual Canopy author, not the saved preference.
- *
- * Users cannot choose a compaction model. Normal compaction reuses the effective model and reasoning settings for context-cache reuse. Only if that model cannot handle the compaction request or produce a summary fitting the target window may the service automatically use another allowed Canopy model with suitable capabilities. This does not change the preference or ordinary answer model, and is permitted even when user model selection is disabled.
- *
- * After compaction, the full next request, including instructions, tools and retained messages, must fit the target's usable prompt budget with room left for an answer, respecting its output limit. If still too large, further compaction attempts are bounded; exhaustion fails with CANNOT_FOLD instead of sending an oversized request or silently discarding history or images. The same checks apply when switching models or falling back. Each compaction entry records its actual author.
- *
- * An empty allowlist or unavailable default does not prevent reading conversations or saving permitted preference changes. A failed Canopy catalogue read is not evidence that a model was removed: use a valid cached snapshot or return MODEL_CATALOG_UNAVAILABLE, without deleting references or inventing capabilities. Catalogue presence is configuration information, not a live provider health check.
- *
- * ## Live output
- *
- * ```
- * GET /api/v1/turns/{turn}/stream?ticket=<ticket>&cursor=<cursor>
- * ```
- *
- * This is a Server-Sent Events stream, which is why it is not in the operation list below and why generated clients have no method for it. Open it with `EventSource`, or any SSE client. It is how live output arrives; polling the document on a timer is not the intended way to follow a conversation.
- *
- * - `path` and `ticket` both come from the conversation document's `stream`. `path` is relative, resolved against the address the document was fetched from;
- * - `cursor` is where to start, from the document's `cursor`. Omit it to start at the beginning of what is still available;
- * - every event is one incremental edit to that same document, so the same rendering logic applies;
- * - the connection has a lifetime and closes cleanly when it ends. With `EventSource` the browser reconnects with `Last-Event-ID` and resumes by itself, so nothing needs handling;
- * - three terminal events: `done` this turn finished (carrying its final status), `paused` the assistant is waiting on the user (fetch the document again and read `wait`), `stalled` execution was interrupted and will not resume — do not reconnect.
- * OpenAPI spec version: 1.0.0
+ * Upload large files with /api/v1/uploads: read /api/v1/upload-configuration first, send exact-sized
+ * parts, then complete the upload and send its attachment id in a message.
+ * OpenAPI spec version: 2.0.0
  */
 import * as zod from "zod";
-export declare const clientActionRequestTimeoutMsMin = 0;
 export declare const clientActionRequestTypeDefault = "function";
+export declare const clientActionRequestTimeoutMsMin = 0;
 export declare const ClientActionRequest: zod.ZodObject<{
+    type: zod.ZodDefault<zod.ZodEnum<{
+        function: "function";
+    }>>;
     function: zod.ZodObject<{
-        description: zod.ZodString;
-        inputSchema: zod.ZodOptional<zod.ZodNullable<zod.ZodRecord<zod.ZodString, zod.ZodUnknown>>>;
         name: zod.ZodString;
+        description: zod.ZodString;
         parameters: zod.ZodOptional<zod.ZodNullable<zod.ZodRecord<zod.ZodString, zod.ZodUnknown>>>;
+        inputSchema: zod.ZodOptional<zod.ZodNullable<zod.ZodRecord<zod.ZodString, zod.ZodUnknown>>>;
     }, zod.z.core.$strict>;
     readOnly: zod.ZodBoolean;
     timeoutMs: zod.ZodOptional<zod.ZodInt>;
-    type: zod.ZodDefault<zod.ZodEnum<{
-        function: "function";
-    }>>;
 }, zod.z.core.$strict>;
 export type ClientActionRequest = zod.input<typeof ClientActionRequest>;
 export type ClientActionRequestOutput = zod.output<typeof ClientActionRequest>;

package/dist/assistant/v1/generated/validators/clientActionRequest.zod.js

@@ -2,68 +2,41 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Assistant API
- * The AI assistant inside the console: conversations, messages, tool calls and their approval, attachments, and connecting the assistant to instant messaging platforms such as WeChat, Feishu and Telegram.
+ * Conversations with agents that operate the Leaflow platform, plus their folders, tasks, files and
+ * chat-app channels.
  *
- * ## A conversation
+ * A conversation is a thread. A person talks to the main thread; the agent in it can start subagents,
+ * which are threads too and run on their own: stopping the main thread does not stop them, and each
+ * subagent tells its parent when it ends. A thread runs in turns; each turn records the model and
+ * reasoning effort it used and its token usage.
  *
- * The path one conversation takes, end to end. It names the operations that path needs, not all of them — the full list is below, and it also covers folders, attachments, memories, skills and the instant messaging side.
+ * Models come from the Canopy catalogue. A new thread uses the model it names, the caller's default
+ * from /api/v1/settings, or the first catalogue model that can run an agent. Reasoning models take one
+ * of three efforts — low, medium or high — which are sent to the model unchanged.
  *
- * 1. `POST /api/v1/threads` creates one;
- * 2. `POST /api/v1/threads/{thread}/messages` sends a message and returns a `turnId` at once; the assistant runs in the background;
- * 3. `GET /api/v1/threads/{thread}` returns the conversation document to render, whose `stream` is where live output arrives;
- * 4. When the assistant needs the user, the document's `wait` says which kind: approving a batch of tool calls (`POST /api/v1/threads/{thread}/approvals/{batch}`), or answering a few questions (`POST /api/v1/threads/{thread}/questions/{item}`). It continues on its own once that is handled.
+ * Read a thread's state with get-thread and its transcript with list-thread-items. Follow both live
+ * over the WebSocket described under /api/v1/realtime/tickets. Slash commands such as /status and
+ * /compact are listed by list-commands and work the same in the console and in chat apps.
  *
- * Sending a file takes two steps: `POST /api/v1/attachments` for an id, then put that id in `attachmentIds` when sending the message. Images, text and arbitrary binaries all go through the same two steps; what differs is how the assistant reads them afterwards, which the upload response's `kind` says.
- *
- * ## Models
- *
- * Canopy owns the model catalogue and capabilities. Model ids in this API, conversation preferences and historical entries are Canopy model ids, with no separate assistant model registration. GET /api/v1/models returns the allowed, usable choices and allowModelSelection. Hide the selector when selection is disabled; the service enforces the same rule.
- *
- * preferredModelId is the user's choice. Null follows the platform default, including later changes. Disabled selection preserves but ignores preferences for new turns. A model removed from the allowlist, retired in Canopy or otherwise unusable falls back to the default without clearing the preference.
- *
- * Users can switch an existing conversation through PATCH /api/v1/threads/{thread} with preferredModelId. This preserves the transcript, starts no turn and makes no inference request. Null returns to following the default. Changes apply from the next turn; a running or waiting turn and messages queued into it retain their captured choices and fallback policy.
- *
- * If the chosen model cannot serve a request, the service may safely fall back to the default without confirmation or an additional switching message. If the default also fails, the turn fails normally; no third model is selected for ordinary answers. Fallback preserves executed tool results and never splices answers from different models. Invalid input, cancellation and a declined tool call do not trigger model fallback. Each entry's model records its actual Canopy author, not the saved preference.
- *
- * Users cannot choose a compaction model. Normal compaction reuses the effective model and reasoning settings for context-cache reuse. Only if that model cannot handle the compaction request or produce a summary fitting the target window may the service automatically use another allowed Canopy model with suitable capabilities. This does not change the preference or ordinary answer model, and is permitted even when user model selection is disabled.
- *
- * After compaction, the full next request, including instructions, tools and retained messages, must fit the target's usable prompt budget with room left for an answer, respecting its output limit. If still too large, further compaction attempts are bounded; exhaustion fails with CANNOT_FOLD instead of sending an oversized request or silently discarding history or images. The same checks apply when switching models or falling back. Each compaction entry records its actual author.
- *
- * An empty allowlist or unavailable default does not prevent reading conversations or saving permitted preference changes. A failed Canopy catalogue read is not evidence that a model was removed: use a valid cached snapshot or return MODEL_CATALOG_UNAVAILABLE, without deleting references or inventing capabilities. Catalogue presence is configuration information, not a live provider health check.
- *
- * ## Live output
- *
- * ```
- * GET /api/v1/turns/{turn}/stream?ticket=<ticket>&cursor=<cursor>
- * ```
- *
- * This is a Server-Sent Events stream, which is why it is not in the operation list below and why generated clients have no method for it. Open it with `EventSource`, or any SSE client. It is how live output arrives; polling the document on a timer is not the intended way to follow a conversation.
- *
- * - `path` and `ticket` both come from the conversation document's `stream`. `path` is relative, resolved against the address the document was fetched from;
- * - `cursor` is where to start, from the document's `cursor`. Omit it to start at the beginning of what is still available;
- * - every event is one incremental edit to that same document, so the same rendering logic applies;
- * - the connection has a lifetime and closes cleanly when it ends. With `EventSource` the browser reconnects with `Last-Event-ID` and resumes by itself, so nothing needs handling;
- * - three terminal events: `done` this turn finished (carrying its final status), `paused` the assistant is waiting on the user (fetch the document again and read `wait`), `stalled` execution was interrupted and will not resume — do not reconnect.
- * OpenAPI spec version: 1.0.0
+ * Upload large files with /api/v1/uploads: read /api/v1/upload-configuration first, send exact-sized
+ * parts, then complete the upload and send its attachment id in a message.
+ * OpenAPI spec version: 2.0.0
  */
 import * as zod from "zod";
 import { ClientFunctionRequest } from "./clientFunctionRequest.zod.js";
-export const clientActionRequestTimeoutMsMin = 0;
 export const clientActionRequestTypeDefault = `function`;
+export const clientActionRequestTimeoutMsMin = 0;
 export const ClientActionRequest = zod
     .strictObject({
+    type: zod.enum(["function"]).default(clientActionRequestTypeDefault),
     function: ClientFunctionRequest,
     readOnly: zod
         .boolean()
-        .describe("True when the action changes nothing outside the client. Anything else goes through this conversation's approval before it runs."),
+        .describe("True when the action changes nothing outside the client. Other actions go through approval before they run."),
     timeoutMs: zod
         .int()
         .min(clientActionRequestTimeoutMsMin)
         .optional()
-        .describe("How long the assistant should wait for this action. Omit for the default; longer values are capped."),
-    type: zod
-        .enum(["function"])
-        .default(clientActionRequestTypeDefault)
-        .describe("Only functions are supported."),
+        .describe("How long the assistant waits for this action. Longer values are capped."),
 })
-    .describe("The shape of a tool in the OpenAI Chat Completions API, plus `readOnly` and `timeoutMs`.");
+    .describe("A tool in the OpenAI Chat Completions format, plus `readOnly` and `timeoutMs`.");

package/dist/assistant/v1/generated/validators/clientContextPart.zod.d.ts

@@ -2,54 +2,30 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Assistant API
- * The AI assistant inside the console: conversations, messages, tool calls and their approval, attachments, and connecting the assistant to instant messaging platforms such as WeChat, Feishu and Telegram.
+ * Conversations with agents that operate the Leaflow platform, plus their folders, tasks, files and
+ * chat-app channels.
  *
- * ## A conversation
+ * A conversation is a thread. A person talks to the main thread; the agent in it can start subagents,
+ * which are threads too and run on their own: stopping the main thread does not stop them, and each
+ * subagent tells its parent when it ends. A thread runs in turns; each turn records the model and
+ * reasoning effort it used and its token usage.
  *
- * The path one conversation takes, end to end. It names the operations that path needs, not all of them — the full list is below, and it also covers folders, attachments, memories, skills and the instant messaging side.
+ * Models come from the Canopy catalogue. A new thread uses the model it names, the caller's default
+ * from /api/v1/settings, or the first catalogue model that can run an agent. Reasoning models take one
+ * of three efforts — low, medium or high — which are sent to the model unchanged.
  *
- * 1. `POST /api/v1/threads` creates one;
- * 2. `POST /api/v1/threads/{thread}/messages` sends a message and returns a `turnId` at once; the assistant runs in the background;
- * 3. `GET /api/v1/threads/{thread}` returns the conversation document to render, whose `stream` is where live output arrives;
- * 4. When the assistant needs the user, the document's `wait` says which kind: approving a batch of tool calls (`POST /api/v1/threads/{thread}/approvals/{batch}`), or answering a few questions (`POST /api/v1/threads/{thread}/questions/{item}`). It continues on its own once that is handled.
+ * Read a thread's state with get-thread and its transcript with list-thread-items. Follow both live
+ * over the WebSocket described under /api/v1/realtime/tickets. Slash commands such as /status and
+ * /compact are listed by list-commands and work the same in the console and in chat apps.
  *
- * Sending a file takes two steps: `POST /api/v1/attachments` for an id, then put that id in `attachmentIds` when sending the message. Images, text and arbitrary binaries all go through the same two steps; what differs is how the assistant reads them afterwards, which the upload response's `kind` says.
- *
- * ## Models
- *
- * Canopy owns the model catalogue and capabilities. Model ids in this API, conversation preferences and historical entries are Canopy model ids, with no separate assistant model registration. GET /api/v1/models returns the allowed, usable choices and allowModelSelection. Hide the selector when selection is disabled; the service enforces the same rule.
- *
- * preferredModelId is the user's choice. Null follows the platform default, including later changes. Disabled selection preserves but ignores preferences for new turns. A model removed from the allowlist, retired in Canopy or otherwise unusable falls back to the default without clearing the preference.
- *
- * Users can switch an existing conversation through PATCH /api/v1/threads/{thread} with preferredModelId. This preserves the transcript, starts no turn and makes no inference request. Null returns to following the default. Changes apply from the next turn; a running or waiting turn and messages queued into it retain their captured choices and fallback policy.
- *
- * If the chosen model cannot serve a request, the service may safely fall back to the default without confirmation or an additional switching message. If the default also fails, the turn fails normally; no third model is selected for ordinary answers. Fallback preserves executed tool results and never splices answers from different models. Invalid input, cancellation and a declined tool call do not trigger model fallback. Each entry's model records its actual Canopy author, not the saved preference.
- *
- * Users cannot choose a compaction model. Normal compaction reuses the effective model and reasoning settings for context-cache reuse. Only if that model cannot handle the compaction request or produce a summary fitting the target window may the service automatically use another allowed Canopy model with suitable capabilities. This does not change the preference or ordinary answer model, and is permitted even when user model selection is disabled.
- *
- * After compaction, the full next request, including instructions, tools and retained messages, must fit the target's usable prompt budget with room left for an answer, respecting its output limit. If still too large, further compaction attempts are bounded; exhaustion fails with CANNOT_FOLD instead of sending an oversized request or silently discarding history or images. The same checks apply when switching models or falling back. Each compaction entry records its actual author.
- *
- * An empty allowlist or unavailable default does not prevent reading conversations or saving permitted preference changes. A failed Canopy catalogue read is not evidence that a model was removed: use a valid cached snapshot or return MODEL_CATALOG_UNAVAILABLE, without deleting references or inventing capabilities. Catalogue presence is configuration information, not a live provider health check.
- *
- * ## Live output
- *
- * ```
- * GET /api/v1/turns/{turn}/stream?ticket=<ticket>&cursor=<cursor>
- * ```
- *
- * This is a Server-Sent Events stream, which is why it is not in the operation list below and why generated clients have no method for it. Open it with `EventSource`, or any SSE client. It is how live output arrives; polling the document on a timer is not the intended way to follow a conversation.
- *
- * - `path` and `ticket` both come from the conversation document's `stream`. `path` is relative, resolved against the address the document was fetched from;
- * - `cursor` is where to start, from the document's `cursor`. Omit it to start at the beginning of what is still available;
- * - every event is one incremental edit to that same document, so the same rendering logic applies;
- * - the connection has a lifetime and closes cleanly when it ends. With `EventSource` the browser reconnects with `Last-Event-ID` and resumes by itself, so nothing needs handling;
- * - three terminal events: `done` this turn finished (carrying its final status), `paused` the assistant is waiting on the user (fetch the document again and read `wait`), `stalled` execution was interrupted and will not resume — do not reconnect.
- * OpenAPI spec version: 1.0.0
+ * Upload large files with /api/v1/uploads: read /api/v1/upload-configuration first, send exact-sized
+ * parts, then complete the upload and send its attachment id in a message.
+ * OpenAPI spec version: 2.0.0
  */
 import * as zod from "zod";
 export declare const ClientContextPart: zod.ZodObject<{
-    data: zod.ZodRecord<zod.ZodString, zod.ZodUnknown>;
     type: zod.ZodString;
+    data: zod.ZodRecord<zod.ZodString, zod.ZodUnknown>;
 }, zod.z.core.$strict>;
 export type ClientContextPart = zod.input<typeof ClientContextPart>;
 export type ClientContextPartOutput = zod.output<typeof ClientContextPart>;

package/dist/assistant/v1/generated/validators/clientContextPart.zod.js

@@ -2,54 +2,30 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Assistant API
- * The AI assistant inside the console: conversations, messages, tool calls and their approval, attachments, and connecting the assistant to instant messaging platforms such as WeChat, Feishu and Telegram.
+ * Conversations with agents that operate the Leaflow platform, plus their folders, tasks, files and
+ * chat-app channels.
  *
- * ## A conversation
+ * A conversation is a thread. A person talks to the main thread; the agent in it can start subagents,
+ * which are threads too and run on their own: stopping the main thread does not stop them, and each
+ * subagent tells its parent when it ends. A thread runs in turns; each turn records the model and
+ * reasoning effort it used and its token usage.
  *
- * The path one conversation takes, end to end. It names the operations that path needs, not all of them — the full list is below, and it also covers folders, attachments, memories, skills and the instant messaging side.
+ * Models come from the Canopy catalogue. A new thread uses the model it names, the caller's default
+ * from /api/v1/settings, or the first catalogue model that can run an agent. Reasoning models take one
+ * of three efforts — low, medium or high — which are sent to the model unchanged.
  *
- * 1. `POST /api/v1/threads` creates one;
- * 2. `POST /api/v1/threads/{thread}/messages` sends a message and returns a `turnId` at once; the assistant runs in the background;
- * 3. `GET /api/v1/threads/{thread}` returns the conversation document to render, whose `stream` is where live output arrives;
- * 4. When the assistant needs the user, the document's `wait` says which kind: approving a batch of tool calls (`POST /api/v1/threads/{thread}/approvals/{batch}`), or answering a few questions (`POST /api/v1/threads/{thread}/questions/{item}`). It continues on its own once that is handled.
+ * Read a thread's state with get-thread and its transcript with list-thread-items. Follow both live
+ * over the WebSocket described under /api/v1/realtime/tickets. Slash commands such as /status and
+ * /compact are listed by list-commands and work the same in the console and in chat apps.
  *
- * Sending a file takes two steps: `POST /api/v1/attachments` for an id, then put that id in `attachmentIds` when sending the message. Images, text and arbitrary binaries all go through the same two steps; what differs is how the assistant reads them afterwards, which the upload response's `kind` says.
- *
- * ## Models
- *
- * Canopy owns the model catalogue and capabilities. Model ids in this API, conversation preferences and historical entries are Canopy model ids, with no separate assistant model registration. GET /api/v1/models returns the allowed, usable choices and allowModelSelection. Hide the selector when selection is disabled; the service enforces the same rule.
- *
- * preferredModelId is the user's choice. Null follows the platform default, including later changes. Disabled selection preserves but ignores preferences for new turns. A model removed from the allowlist, retired in Canopy or otherwise unusable falls back to the default without clearing the preference.
- *
- * Users can switch an existing conversation through PATCH /api/v1/threads/{thread} with preferredModelId. This preserves the transcript, starts no turn and makes no inference request. Null returns to following the default. Changes apply from the next turn; a running or waiting turn and messages queued into it retain their captured choices and fallback policy.
- *
- * If the chosen model cannot serve a request, the service may safely fall back to the default without confirmation or an additional switching message. If the default also fails, the turn fails normally; no third model is selected for ordinary answers. Fallback preserves executed tool results and never splices answers from different models. Invalid input, cancellation and a declined tool call do not trigger model fallback. Each entry's model records its actual Canopy author, not the saved preference.
- *
- * Users cannot choose a compaction model. Normal compaction reuses the effective model and reasoning settings for context-cache reuse. Only if that model cannot handle the compaction request or produce a summary fitting the target window may the service automatically use another allowed Canopy model with suitable capabilities. This does not change the preference or ordinary answer model, and is permitted even when user model selection is disabled.
- *
- * After compaction, the full next request, including instructions, tools and retained messages, must fit the target's usable prompt budget with room left for an answer, respecting its output limit. If still too large, further compaction attempts are bounded; exhaustion fails with CANNOT_FOLD instead of sending an oversized request or silently discarding history or images. The same checks apply when switching models or falling back. Each compaction entry records its actual author.
- *
- * An empty allowlist or unavailable default does not prevent reading conversations or saving permitted preference changes. A failed Canopy catalogue read is not evidence that a model was removed: use a valid cached snapshot or return MODEL_CATALOG_UNAVAILABLE, without deleting references or inventing capabilities. Catalogue presence is configuration information, not a live provider health check.
- *
- * ## Live output
- *
- * ```
- * GET /api/v1/turns/{turn}/stream?ticket=<ticket>&cursor=<cursor>
- * ```
- *
- * This is a Server-Sent Events stream, which is why it is not in the operation list below and why generated clients have no method for it. Open it with `EventSource`, or any SSE client. It is how live output arrives; polling the document on a timer is not the intended way to follow a conversation.
- *
- * - `path` and `ticket` both come from the conversation document's `stream`. `path` is relative, resolved against the address the document was fetched from;
- * - `cursor` is where to start, from the document's `cursor`. Omit it to start at the beginning of what is still available;
- * - every event is one incremental edit to that same document, so the same rendering logic applies;
- * - the connection has a lifetime and closes cleanly when it ends. With `EventSource` the browser reconnects with `Last-Event-ID` and resumes by itself, so nothing needs handling;
- * - three terminal events: `done` this turn finished (carrying its final status), `paused` the assistant is waiting on the user (fetch the document again and read `wait`), `stalled` execution was interrupted and will not resume — do not reconnect.
- * OpenAPI spec version: 1.0.0
+ * Upload large files with /api/v1/uploads: read /api/v1/upload-configuration first, send exact-sized
+ * parts, then complete the upload and send its attachment id in a message.
+ * OpenAPI spec version: 2.0.0
  */
 import * as zod from "zod";
 export const ClientContextPart = zod
     .strictObject({
-    data: zod.record(zod.string(), zod.unknown()),
     type: zod.string(),
+    data: zod.record(zod.string(), zod.unknown()),
 })
-    .describe("A `data-*` part as it was sent.");
+    .describe("A `data-*` part as the client sent it.");

package/dist/assistant/v1/generated/validators/clientContextRequest.zod.d.ts

@@ -2,70 +2,46 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Assistant API
- * The AI assistant inside the console: conversations, messages, tool calls and their approval, attachments, and connecting the assistant to instant messaging platforms such as WeChat, Feishu and Telegram.
+ * Conversations with agents that operate the Leaflow platform, plus their folders, tasks, files and
+ * chat-app channels.
  *
- * ## A conversation
+ * A conversation is a thread. A person talks to the main thread; the agent in it can start subagents,
+ * which are threads too and run on their own: stopping the main thread does not stop them, and each
+ * subagent tells its parent when it ends. A thread runs in turns; each turn records the model and
+ * reasoning effort it used and its token usage.
  *
- * The path one conversation takes, end to end. It names the operations that path needs, not all of them — the full list is below, and it also covers folders, attachments, memories, skills and the instant messaging side.
+ * Models come from the Canopy catalogue. A new thread uses the model it names, the caller's default
+ * from /api/v1/settings, or the first catalogue model that can run an agent. Reasoning models take one
+ * of three efforts — low, medium or high — which are sent to the model unchanged.
  *
- * 1. `POST /api/v1/threads` creates one;
- * 2. `POST /api/v1/threads/{thread}/messages` sends a message and returns a `turnId` at once; the assistant runs in the background;
- * 3. `GET /api/v1/threads/{thread}` returns the conversation document to render, whose `stream` is where live output arrives;
- * 4. When the assistant needs the user, the document's `wait` says which kind: approving a batch of tool calls (`POST /api/v1/threads/{thread}/approvals/{batch}`), or answering a few questions (`POST /api/v1/threads/{thread}/questions/{item}`). It continues on its own once that is handled.
+ * Read a thread's state with get-thread and its transcript with list-thread-items. Follow both live
+ * over the WebSocket described under /api/v1/realtime/tickets. Slash commands such as /status and
+ * /compact are listed by list-commands and work the same in the console and in chat apps.
  *
- * Sending a file takes two steps: `POST /api/v1/attachments` for an id, then put that id in `attachmentIds` when sending the message. Images, text and arbitrary binaries all go through the same two steps; what differs is how the assistant reads them afterwards, which the upload response's `kind` says.
- *
- * ## Models
- *
- * Canopy owns the model catalogue and capabilities. Model ids in this API, conversation preferences and historical entries are Canopy model ids, with no separate assistant model registration. GET /api/v1/models returns the allowed, usable choices and allowModelSelection. Hide the selector when selection is disabled; the service enforces the same rule.
- *
- * preferredModelId is the user's choice. Null follows the platform default, including later changes. Disabled selection preserves but ignores preferences for new turns. A model removed from the allowlist, retired in Canopy or otherwise unusable falls back to the default without clearing the preference.
- *
- * Users can switch an existing conversation through PATCH /api/v1/threads/{thread} with preferredModelId. This preserves the transcript, starts no turn and makes no inference request. Null returns to following the default. Changes apply from the next turn; a running or waiting turn and messages queued into it retain their captured choices and fallback policy.
- *
- * If the chosen model cannot serve a request, the service may safely fall back to the default without confirmation or an additional switching message. If the default also fails, the turn fails normally; no third model is selected for ordinary answers. Fallback preserves executed tool results and never splices answers from different models. Invalid input, cancellation and a declined tool call do not trigger model fallback. Each entry's model records its actual Canopy author, not the saved preference.
- *
- * Users cannot choose a compaction model. Normal compaction reuses the effective model and reasoning settings for context-cache reuse. Only if that model cannot handle the compaction request or produce a summary fitting the target window may the service automatically use another allowed Canopy model with suitable capabilities. This does not change the preference or ordinary answer model, and is permitted even when user model selection is disabled.
- *
- * After compaction, the full next request, including instructions, tools and retained messages, must fit the target's usable prompt budget with room left for an answer, respecting its output limit. If still too large, further compaction attempts are bounded; exhaustion fails with CANNOT_FOLD instead of sending an oversized request or silently discarding history or images. The same checks apply when switching models or falling back. Each compaction entry records its actual author.
- *
- * An empty allowlist or unavailable default does not prevent reading conversations or saving permitted preference changes. A failed Canopy catalogue read is not evidence that a model was removed: use a valid cached snapshot or return MODEL_CATALOG_UNAVAILABLE, without deleting references or inventing capabilities. Catalogue presence is configuration information, not a live provider health check.
- *
- * ## Live output
- *
- * ```
- * GET /api/v1/turns/{turn}/stream?ticket=<ticket>&cursor=<cursor>
- * ```
- *
- * This is a Server-Sent Events stream, which is why it is not in the operation list below and why generated clients have no method for it. Open it with `EventSource`, or any SSE client. It is how live output arrives; polling the document on a timer is not the intended way to follow a conversation.
- *
- * - `path` and `ticket` both come from the conversation document's `stream`. `path` is relative, resolved against the address the document was fetched from;
- * - `cursor` is where to start, from the document's `cursor`. Omit it to start at the beginning of what is still available;
- * - every event is one incremental edit to that same document, so the same rendering logic applies;
- * - the connection has a lifetime and closes cleanly when it ends. With `EventSource` the browser reconnects with `Last-Event-ID` and resumes by itself, so nothing needs handling;
- * - three terminal events: `done` this turn finished (carrying its final status), `paused` the assistant is waiting on the user (fetch the document again and read `wait`), `stalled` execution was interrupted and will not resume — do not reconnect.
- * OpenAPI spec version: 1.0.0
+ * Upload large files with /api/v1/uploads: read /api/v1/upload-configuration first, send exact-sized
+ * parts, then complete the upload and send its attachment id in a message.
+ * OpenAPI spec version: 2.0.0
  */
 import * as zod from "zod";
-export declare const clientContextRequestActionsMax = 32;
 export declare const clientContextRequestClientIdMax = 128;
 export declare const clientContextRequestLabelMax = 64;
+export declare const clientContextRequestActionsMax = 32;
 export declare const ClientContextRequest: zod.ZodObject<{
+    clientId: zod.ZodString;
+    label: zod.ZodOptional<zod.ZodString>;
     actions: zod.ZodOptional<zod.ZodNullable<zod.ZodArray<zod.ZodObject<{
+        type: zod.ZodDefault<zod.ZodEnum<{
+            function: "function";
+        }>>;
         function: zod.ZodObject<{
-            description: zod.ZodString;
-            inputSchema: zod.ZodOptional<zod.ZodNullable<zod.ZodRecord<zod.ZodString, zod.ZodUnknown>>>;
             name: zod.ZodString;
+            description: zod.ZodString;
             parameters: zod.ZodOptional<zod.ZodNullable<zod.ZodRecord<zod.ZodString, zod.ZodUnknown>>>;
+            inputSchema: zod.ZodOptional<zod.ZodNullable<zod.ZodRecord<zod.ZodString, zod.ZodUnknown>>>;
         }, zod.z.core.$strict>;
         readOnly: zod.ZodBoolean;
         timeoutMs: zod.ZodOptional<zod.ZodInt>;
-        type: zod.ZodDefault<zod.ZodEnum<{
-            function: "function";
-        }>>;
     }, zod.z.core.$strict>>>>;
-    clientId: zod.ZodString;
-    label: zod.ZodOptional<zod.ZodString>;
 }, zod.z.core.$strict>;
 export type ClientContextRequest = zod.input<typeof ClientContextRequest>;
 export type ClientContextRequestOutput = zod.output<typeof ClientContextRequest>;

package/dist/assistant/v1/generated/validators/clientContextRequest.zod.js

@@ -2,71 +2,47 @@
  * Generated by orval 🍺
  * Do not edit manually.
  * Leaflow Assistant API
- * The AI assistant inside the console: conversations, messages, tool calls and their approval, attachments, and connecting the assistant to instant messaging platforms such as WeChat, Feishu and Telegram.
+ * Conversations with agents that operate the Leaflow platform, plus their folders, tasks, files and
+ * chat-app channels.
  *
- * ## A conversation
+ * A conversation is a thread. A person talks to the main thread; the agent in it can start subagents,
+ * which are threads too and run on their own: stopping the main thread does not stop them, and each
+ * subagent tells its parent when it ends. A thread runs in turns; each turn records the model and
+ * reasoning effort it used and its token usage.
  *
- * The path one conversation takes, end to end. It names the operations that path needs, not all of them — the full list is below, and it also covers folders, attachments, memories, skills and the instant messaging side.
+ * Models come from the Canopy catalogue. A new thread uses the model it names, the caller's default
+ * from /api/v1/settings, or the first catalogue model that can run an agent. Reasoning models take one
+ * of three efforts — low, medium or high — which are sent to the model unchanged.
  *
- * 1. `POST /api/v1/threads` creates one;
- * 2. `POST /api/v1/threads/{thread}/messages` sends a message and returns a `turnId` at once; the assistant runs in the background;
- * 3. `GET /api/v1/threads/{thread}` returns the conversation document to render, whose `stream` is where live output arrives;
- * 4. When the assistant needs the user, the document's `wait` says which kind: approving a batch of tool calls (`POST /api/v1/threads/{thread}/approvals/{batch}`), or answering a few questions (`POST /api/v1/threads/{thread}/questions/{item}`). It continues on its own once that is handled.
+ * Read a thread's state with get-thread and its transcript with list-thread-items. Follow both live
+ * over the WebSocket described under /api/v1/realtime/tickets. Slash commands such as /status and
+ * /compact are listed by list-commands and work the same in the console and in chat apps.
  *
- * Sending a file takes two steps: `POST /api/v1/attachments` for an id, then put that id in `attachmentIds` when sending the message. Images, text and arbitrary binaries all go through the same two steps; what differs is how the assistant reads them afterwards, which the upload response's `kind` says.
- *
- * ## Models
- *
- * Canopy owns the model catalogue and capabilities. Model ids in this API, conversation preferences and historical entries are Canopy model ids, with no separate assistant model registration. GET /api/v1/models returns the allowed, usable choices and allowModelSelection. Hide the selector when selection is disabled; the service enforces the same rule.
- *
- * preferredModelId is the user's choice. Null follows the platform default, including later changes. Disabled selection preserves but ignores preferences for new turns. A model removed from the allowlist, retired in Canopy or otherwise unusable falls back to the default without clearing the preference.
- *
- * Users can switch an existing conversation through PATCH /api/v1/threads/{thread} with preferredModelId. This preserves the transcript, starts no turn and makes no inference request. Null returns to following the default. Changes apply from the next turn; a running or waiting turn and messages queued into it retain their captured choices and fallback policy.
- *
- * If the chosen model cannot serve a request, the service may safely fall back to the default without confirmation or an additional switching message. If the default also fails, the turn fails normally; no third model is selected for ordinary answers. Fallback preserves executed tool results and never splices answers from different models. Invalid input, cancellation and a declined tool call do not trigger model fallback. Each entry's model records its actual Canopy author, not the saved preference.
- *
- * Users cannot choose a compaction model. Normal compaction reuses the effective model and reasoning settings for context-cache reuse. Only if that model cannot handle the compaction request or produce a summary fitting the target window may the service automatically use another allowed Canopy model with suitable capabilities. This does not change the preference or ordinary answer model, and is permitted even when user model selection is disabled.
- *
- * After compaction, the full next request, including instructions, tools and retained messages, must fit the target's usable prompt budget with room left for an answer, respecting its output limit. If still too large, further compaction attempts are bounded; exhaustion fails with CANNOT_FOLD instead of sending an oversized request or silently discarding history or images. The same checks apply when switching models or falling back. Each compaction entry records its actual author.
- *
- * An empty allowlist or unavailable default does not prevent reading conversations or saving permitted preference changes. A failed Canopy catalogue read is not evidence that a model was removed: use a valid cached snapshot or return MODEL_CATALOG_UNAVAILABLE, without deleting references or inventing capabilities. Catalogue presence is configuration information, not a live provider health check.
- *
- * ## Live output
- *
- * ```
- * GET /api/v1/turns/{turn}/stream?ticket=<ticket>&cursor=<cursor>
- * ```
- *
- * This is a Server-Sent Events stream, which is why it is not in the operation list below and why generated clients have no method for it. Open it with `EventSource`, or any SSE client. It is how live output arrives; polling the document on a timer is not the intended way to follow a conversation.
- *
- * - `path` and `ticket` both come from the conversation document's `stream`. `path` is relative, resolved against the address the document was fetched from;
- * - `cursor` is where to start, from the document's `cursor`. Omit it to start at the beginning of what is still available;
- * - every event is one incremental edit to that same document, so the same rendering logic applies;
- * - the connection has a lifetime and closes cleanly when it ends. With `EventSource` the browser reconnects with `Last-Event-ID` and resumes by itself, so nothing needs handling;
- * - three terminal events: `done` this turn finished (carrying its final status), `paused` the assistant is waiting on the user (fetch the document again and read `wait`), `stalled` execution was interrupted and will not resume — do not reconnect.
- * OpenAPI spec version: 1.0.0
+ * Upload large files with /api/v1/uploads: read /api/v1/upload-configuration first, send exact-sized
+ * parts, then complete the upload and send its attachment id in a message.
+ * OpenAPI spec version: 2.0.0
  */
 import * as zod from "zod";
 import { ClientActionRequest } from "./clientActionRequest.zod.js";
-export const clientContextRequestActionsMax = 32;
 export const clientContextRequestClientIdMax = 128;
 export const clientContextRequestLabelMax = 64;
+export const clientContextRequestActionsMax = 32;
 export const ClientContextRequest = zod
     .strictObject({
-    actions: zod
-        .array(ClientActionRequest)
-        .max(clientContextRequestActionsMax)
-        .nullish()
-        .describe("The actions on offer right now.\n\nOmit it when nothing changed since the last message. Send `[]` to say there is nothing\non offer — those are different: a client that moves off the page it was attached to has\nto be able to say so, or the assistant keeps calling actions that no longer make sense."),
     clientId: zod
         .string()
         .min(1)
         .max(clientContextRequestClientIdMax)
-        .describe("Identifies this client while it stays open. Any stable string; one per tab or process."),
+        .describe("Identifies this client while it stays open; one per tab or process."),
     label: zod
         .string()
         .max(clientContextRequestLabelMax)
         .optional()
-        .describe('How the client calls itself, for example "Leaflow console (web)".'),
+        .describe('How the client names itself, for example "Leaflow console (web)".'),
+    actions: zod
+        .array(ClientActionRequest)
+        .max(clientContextRequestActionsMax)
+        .nullish()
+        .describe("The actions on offer now. `[]` withdraws all of them; null keeps the previous declaration."),
 })
-    .describe("Which client this is and what it can do, so the assistant can ask it to do those things\nwhile it answers. What the operator is looking at travels as `data-*` parts on the message.\n\nSend `actions` only when they differ from the last message on this conversation. Sending the\nblock without `actions` keeps whatever was declared before.");
+    .describe("Which client this is and which actions it offers, so the assistant can ask it to run them while\nit answers. Send `actions` only when they changed since the last message; omitting them keeps\nwhat was declared before.");