@dalmia/calibrate-mcp 0.0.15 → 0.0.16

package/bin/mcp-server.js

@@ -51392,9 +51392,9 @@ var init_config = __esm(() => {
   SDK_METADATA = {
     language: "typescript",
     openapiDocVersion: "0.1.0",
-    sdkVersion: "0.0.15",
+    sdkVersion: "0.0.16",
     genVersion: "2.915.1",
-    userAgent: "speakeasy-sdk/mcp-typescript 0.0.15 2.915.1 0.1.0 @dalmia/calibrate-mcp"
+    userAgent: "speakeasy-sdk/mcp-typescript 0.0.16 2.915.1 0.1.0 @dalmia/calibrate-mcp"
   };
 });
 
@@ -55202,7 +55202,7 @@ var init_annotationitempayload = __esm(() => {
   init_zod();
   AnnotationItemPayload$zodSchema = object({
     annotations: record(string2(), any()).nullable().optional().describe("Human annotations to seed, keyed by evaluator ID. Each evaluator ID must be linked to the task. Put the judgement in `value`, a bool for binary or a number for rating, with optional `reasoning`. Requires `annotator_id`"),
-    payload: any().optional().describe("The item's payload. Its shape depends on the task `type`. `payload['name']` is **required** for every type and must be unique within the task")
+    payload: any().optional().describe("The item's payload. `name` is required and unique within the task for every type. The other fields depend on the task `type`:\n\n- `stt`: `name`, `reference_transcript`, `predicted_transcript`\n- `llm`: `name`, `chat_history` (list of `{role, content}` turns ending at the user turn), `agent_response` (the reply to judge), optional `evaluator_variables`\n- `llm-general`: `name`, `input`, `output`, optional `evaluator_variables`\n- `conversation`: `name`, `transcript` (list of `{role, content}` turns), optional `evaluator_variables`\n\n`evaluator_variables` maps an evaluator ID to that evaluator's `{{\"{{\"}}variable}}` values for this item")
   });
 });
 
@@ -55359,16 +55359,15 @@ var init_annotationtaskcreate = __esm(() => {
   init_zod();
   AnnotationTaskCreateType$zodSchema = _enum([
     "stt",
-    "tts",
     "llm",
     "llm-general",
     "conversation"
-  ]).describe("Task type. Determines the shape of each item's payload.\n- `stt`: judge a transcript on its own\n- `tts`: judge synthesized speech output\n- `llm`: judge one response with its conversation history\n- `llm-general`: judge a standalone `input -> output` pair\n- `conversation`: judge a full conversation");
+  ]).describe("Task type. Determines the shape of each item's payload.\n- `stt`: judge a transcript on its own\n- `llm`: judge one response with its conversation history\n- `llm-general`: judge a standalone `input -> output` pair\n- `conversation`: judge a full conversation");
   AnnotationTaskCreate$zodSchema = object({
     description: string2().nullable().optional().describe("A description for the task. Omit for none"),
     evaluator_ids: array(string2()).nullable().optional().describe("IDs of evaluators to link when the task is created, in order. Each must be one you created or a built-in default. Omit to create with no linked evaluators"),
     name: string2().describe("Task name, unique within your workspace"),
-    type: AnnotationTaskCreateType$zodSchema.describe("Task type. Determines the shape of each item's payload.\n- `stt`: judge a transcript on its own\n- `tts`: judge synthesized speech output\n- `llm`: judge one response with its conversation history\n- `llm-general`: judge a standalone `input -> output` pair\n- `conversation`: judge a full conversation")
+    type: AnnotationTaskCreateType$zodSchema.describe("Task type. Determines the shape of each item's payload.\n- `stt`: judge a transcript on its own\n- `llm`: judge one response with its conversation history\n- `llm-general`: judge a standalone `input -> output` pair\n- `conversation`: judge a full conversation")
   });
 });
 
@@ -55692,11 +55691,10 @@ var init_annotationtaskresponse = __esm(() => {
   init_zod();
   AnnotationTaskResponseType$zodSchema = _enum([
     "stt",
-    "tts",
     "llm",
     "llm-general",
     "conversation"
-  ]).describe("Task type. Determines the shape of each item's payload.\n- `stt`: judge a transcript on its own\n- `tts`: judge synthesized speech output\n- `llm`: judge one response with its conversation history\n- `llm-general`: judge a standalone `input -> output` pair\n- `conversation`: judge a full conversation");
+  ]).describe("Task type. Determines the shape of each item's payload.\n- `stt`: judge a transcript on its own\n- `llm`: judge one response with its conversation history\n- `llm-general`: judge a standalone `input -> output` pair\n- `conversation`: judge a full conversation");
   AnnotationTaskResponse$zodSchema = object({
     created_at: string2().describe("When the task was created (ISO 8601 UTC)"),
     description: string2().nullable().optional().describe("The task's description"),
@@ -55705,7 +55703,7 @@ var init_annotationtaskresponse = __esm(() => {
     items: array(record(string2(), any())).optional().describe("The task's items, each with its agreement stats. You get these when you fetch one task by ID, not when you list tasks"),
     jobs: array(record(string2(), any())).optional().describe("The task's labelling jobs. You get these when you fetch one task by ID, not when you list tasks"),
     name: string2().describe("Name of the task"),
-    type: AnnotationTaskResponseType$zodSchema.describe("Task type. Determines the shape of each item's payload.\n- `stt`: judge a transcript on its own\n- `tts`: judge synthesized speech output\n- `llm`: judge one response with its conversation history\n- `llm-general`: judge a standalone `input -> output` pair\n- `conversation`: judge a full conversation"),
+    type: AnnotationTaskResponseType$zodSchema.describe("Task type. Determines the shape of each item's payload.\n- `stt`: judge a transcript on its own\n- `llm`: judge one response with its conversation history\n- `llm-general`: judge a standalone `input -> output` pair\n- `conversation`: judge a full conversation"),
     updated_at: string2().describe("When the task was last updated (ISO 8601 UTC)"),
     uuid: string2().describe("Unique ID for the task")
   });
@@ -56245,7 +56243,6 @@ var init_tasksummaryresponse = __esm(() => {
   init_paginationmeta();
   TaskType$zodSchema = _enum([
     "stt",
-    "tts",
     "llm",
     "llm-general",
     "conversation"
@@ -58459,7 +58456,7 @@ fields to change (name, config). Returns the updated test. Fails with
 function createMCPServer(deps) {
   const server = new McpServer({
     name: "CalibrateMcp",
-    version: "0.0.15"
+    version: "0.0.16"
   });
   const getClient = deps.getSDK || (() => new CalibrateMcpCore({
     security: deps.security,
@@ -59665,7 +59662,7 @@ http_headers = { "api-key-auth" = "YOUR_API_KEY_AUTH" }`;
         <h1>Instructions</h1>
         <p>One-click installation for Claude Desktop users</p>
         <div class="instruction-item">
-          <a href="https://github.com/dalmia/calibrate-mcp/releases/download/v0.0.15/mcp-server.mcpb" download="mcp-server.mcpb" class="action-button header-action" style="display: inline-flex; margin-bottom: 16px;">
+          <a href="https://github.com/dalmia/calibrate-mcp/releases/download/v0.0.16/mcp-server.mcpb" download="mcp-server.mcpb" class="action-button header-action" style="display: inline-flex; margin-bottom: 16px;">
             \uD83D\uDCE5 Download MCP Bundle
           </a>
         </div>
@@ -62317,7 +62314,7 @@ var routes = buildRouteMap({
 var app = buildApplication(routes, {
   name: "mcp",
   versionInfo: {
-    currentVersion: "0.0.15"
+    currentVersion: "0.0.16"
   }
 });
 run(app, process4.argv.slice(2), buildContext(process4));
@@ -62325,5 +62322,5 @@ export {
   app
 };
 
-//# debugId=95D5AEF339D84AE864756E2164756E21
+//# debugId=2A3531F2223C640064756E2164756E21
 //# sourceMappingURL=mcp-server.js.map