@arizeai/phoenix-client 6.14.2 → 7.0.0

package/docs/experiments.mdx

@@ -111,11 +111,18 @@ This pattern is useful when:
 
 ## Model-Backed Example
 
-If you want a model-backed experiment with automatic tracing and an LLM-as-a-judge evaluator, this is the core pattern:
+If you want a model-backed experiment with automatic tracing and an LLM-as-a-judge evaluator, this is the core pattern.
+
+Install the AI SDK packages alongside the Phoenix packages — `@ai-sdk/otel` provides the telemetry integration that sends AI SDK traces to Phoenix:
+
+```bash
+npm install @arizeai/phoenix-client @arizeai/phoenix-evals ai @ai-sdk/openai @ai-sdk/otel
+```
 
 ```ts
 import { openai } from "@ai-sdk/openai";
-import { createOrGetDataset } from "@arizeai/phoenix-client/datasets";
+import { OpenTelemetry } from "@ai-sdk/otel";
+import { createDataset } from "@arizeai/phoenix-client/datasets";
 import { runExperiment } from "@arizeai/phoenix-client/experiments";
 import type { ExperimentTask } from "@arizeai/phoenix-client/types/experiments";
 import { createClassificationEvaluator } from "@arizeai/phoenix-evals";
@@ -135,7 +142,7 @@ const main = async () => {
     },
   });
 
-  const dataset = await createOrGetDataset({
+  const dataset = await createDataset({
     name: "correctness-eval",
     description: "Evaluate the correctness of the model",
     examples: [
@@ -157,23 +164,15 @@ const main = async () => {
       throw new Error("Invalid input: context must be a string");
     }
 
+    // The per-call `@ai-sdk/otel` integration traces this call through the
+    // experiment's tracer provider that runExperiment mounts while tasks run.
     return generateText({
       model,
-      experimental_telemetry: {
-        isEnabled: true,
-      },
-      prompt: [
-        {
-          role: "system",
-          content: `You answer questions based on this context: ${example.input.context}`,
-        },
-        {
-          role: "user",
-          content: example.input.question,
-        },
-      ],
+      instructions: `You answer questions based on this context: ${example.input.context}`,
+      prompt: example.input.question,
+      telemetry: { integrations: [new OpenTelemetry()] },
     }).then((response) => {
-      if (response.text) {
+      if (typeof response.text === "string") {
         return response.text;
       }
       throw new Error("Invalid response: text is required");
@@ -200,9 +199,9 @@ main().catch(console.error);
 
 ## What This Example Shows
 
-- `createOrGetDataset()` creates or reuses the dataset the experiment will run against
+- `createDataset()` creates or reuses the dataset the experiment will run against (it upserts by name)
 - `task` receives the full dataset example object
-- `generateText()` emits traces that Phoenix can attach to the experiment when telemetry is enabled
+- `generateText()` emits traces that Phoenix attaches to the experiment because the call passes the `@ai-sdk/otel` integration via `telemetry.integrations` — without it, no AI SDK spans reach Phoenix
 - `createClassificationEvaluator()` from `@arizeai/phoenix-evals` can be passed directly to `runExperiment()`
 - `runExperiment()` records both task runs and evaluation runs in Phoenix
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@arizeai/phoenix-client",
-  "version": "6.14.2",
+  "version": "7.0.0",
   "description": "A client for the Phoenix API",
   "keywords": [
     "arize",
@@ -96,34 +96,34 @@
   },
   "dependencies": {
     "@arizeai/openinference-semantic-conventions": "^2.5.0",
-    "@arizeai/openinference-vercel": "^2.8.1",
     "async": "^3.2.6",
     "openapi-fetch": "^0.17.0",
     "tiny-invariant": "^1.3.3",
     "zod": "^4.4.3",
-    "@arizeai/phoenix-otel": "2.1.0",
-    "@arizeai/phoenix-config": "0.4.0"
+    "@arizeai/phoenix-config": "0.4.0",
+    "@arizeai/phoenix-otel": "2.1.0"
   },
   "devDependencies": {
-    "@ai-sdk/openai": "^3.0.86",
+    "@ai-sdk/openai": "^4.0.0",
+    "@ai-sdk/otel": "^1.0.0",
     "@anthropic-ai/sdk": "^0.111.0",
     "@opentelemetry/api": "^1.9.1",
     "@opentelemetry/sdk-trace-node": "^2.9.0",
     "@types/async": "^3.2.25",
     "@types/node": "^26.1.1",
-    "ai": "^6.0.230",
+    "ai": "^7.0.0",
     "dotenv": "^17.4.2",
     "jest": "^30.4.2",
     "openai": "^6.48.0",
     "openapi-typescript": "^7.13.0",
     "tsx": "^4.23.1",
     "vitest": "^4.1.10",
-    "@arizeai/phoenix-testing": "0.0.0",
-    "@arizeai/phoenix-evals": "1.2.0"
+    "@arizeai/phoenix-evals": "2.0.0",
+    "@arizeai/phoenix-testing": "0.0.0"
   },
   "peerDependencies": {
     "@anthropic-ai/sdk": "^0.35.0",
-    "ai": "^6.0.90",
+    "ai": "^7.0.0",
     "jest": ">=27",
     "openai": "^6.10.0",
     "vitest": ">=1"
@@ -156,6 +156,6 @@
     "prebuild": "pnpm run clean && pnpm run generate",
     "test": "vitest run",
     "test:watch": "vitest watch",
-    "typecheck": "tsc --noEmit"
+    "typecheck": "tsc --noEmit && tsc -p examples/tsconfig.examples.json"
   }
 }

package/src/prompts/sdks/toAI.ts

@@ -1,4 +1,6 @@
-import type { ModelMessage, ToolChoice, ToolSet } from "ai";
+import type { ModelMessage, ToolChoice, ToolSet } from "ai" with {
+  "resolution-mode": "import",
+};
 import invariant from "tiny-invariant";
 
 import {