How to use StreamV2 with @mastra/client-js using the aisdk format?
My goal is to return:
The mastra docs show this:
BUT with the client-js,
* the streamVNext does not have a stream.fullStream
* it also does not have toUImessageStreamResponse()
So I am doing this instead:
Am I missing an equivalent to fullStream / toUIMessageStreamResponse() in client-js, or is processDataStream the intended path? Any pointers appreciated. Thanks!
Thanks everyone! I really appreciate your help.
return stream.toUIMessageStreamResponse();return stream.toUIMessageStreamResponse();The mastra docs show this:
const stream = await agent.streamVNext("Tell me a story", {
format: 'aisdk',
stopWhen: stepCountIs(3), // Stop after 3 steps
modelSettings: {
temperature: 0.7,
},
});
// Use with AI SDK v5 compatible interfaces
for await (const part of stream.fullStream) {
if (part.type === 'text-delta') {
console.log(part.text);
}
}
// In an API route for frontend integration
return stream.toUIMessageStreamResponse();const stream = await agent.streamVNext("Tell me a story", {
format: 'aisdk',
stopWhen: stepCountIs(3), // Stop after 3 steps
modelSettings: {
temperature: 0.7,
},
});
// Use with AI SDK v5 compatible interfaces
for await (const part of stream.fullStream) {
if (part.type === 'text-delta') {
console.log(part.text);
}
}
// In an API route for frontend integration
return stream.toUIMessageStreamResponse();BUT with the client-js,
* the streamVNext does not have a stream.fullStream
* it also does not have toUImessageStreamResponse()
So I am doing this instead:
myAgent = mastraClient.getAgent("agent");
mastraResponse = myagent.streamVNext({
})
// Simplified
mastraResponse.processDataStream({
onChunk: async (chunk) => {
try {
// Extract text content from chunk based on its type
if (chunk.type === "text-delta" && "text" in chunk.payload) {
console.log("📝 Text delta:", chunk.payload.text);
// Don't write to process.stdout in web API context - just log
}
if (chunk.type === "tool-call" && "toolName" in chunk.payload) {
// Don't write to process.stdout in web API context - just log
}
// Log other chunk types for debugging
if (chunk.type === "response-metadata") {
console.log("📊 Response metadata received");
}
} catch (error) {
console.error("❌ Error processing chunk:", error);
// Don't log the entire chunk as it might be too large or cause circular reference issues
}
},myAgent = mastraClient.getAgent("agent");
mastraResponse = myagent.streamVNext({
})
// Simplified
mastraResponse.processDataStream({
onChunk: async (chunk) => {
try {
// Extract text content from chunk based on its type
if (chunk.type === "text-delta" && "text" in chunk.payload) {
console.log("📝 Text delta:", chunk.payload.text);
// Don't write to process.stdout in web API context - just log
}
if (chunk.type === "tool-call" && "toolName" in chunk.payload) {
// Don't write to process.stdout in web API context - just log
}
// Log other chunk types for debugging
if (chunk.type === "response-metadata") {
console.log("📊 Response metadata received");
}
} catch (error) {
console.error("❌ Error processing chunk:", error);
// Don't log the entire chunk as it might be too large or cause circular reference issues
}
},Am I missing an equivalent to fullStream / toUIMessageStreamResponse() in client-js, or is processDataStream the intended path? Any pointers appreciated. Thanks!
Thanks everyone! I really appreciate your help.
