fix: ensure output for non-streaming models (#369)

Co-authored-by: mneves75 <mneves75@users.noreply.github.com>

CHANGELOG.md

@@ -24,6 +24,7 @@
 - Gateway/CLI: add daemon runtime selection (Node recommended; Bun optional) and document WhatsApp/Baileys Bun WebSocket instability on reconnect.
 - CLI: add `clawdbot docs` live docs search with pretty output.
 - Agent: treat compaction retry AbortError as a fallback trigger without swallowing non-abort errors. Thanks @erikpr1994 for PR #341.
+- Agent: deliver final replies for non-streaming models when block chunking is enabled. Thanks @mneves75 for PR #369.
 - Sub-agents: allow `sessions_spawn` model overrides and error on invalid models. Thanks @azade-c for PR #298.
 - Sub-agents: skip invalid model overrides with a warning and keep the run alive; tool exceptions now return tool errors instead of crashing the agent.
 - Heartbeat: default interval 30m; clarified default prompt usage and HEARTBEAT.md template behavior.

src/agents/pi-embedded-subscribe.test.ts

@@ -400,6 +400,40 @@ describe("subscribeEmbeddedPiSession", () => {
     expect(subscription.assistantTexts).toEqual(["Hello world"]);
   });
 
+  it("populates assistantTexts for non-streaming models with chunking enabled", () => {
+    // Non-streaming models (e.g. zai/glm-4.7): no text_delta events; message_end
+    // must still populate assistantTexts so providers can deliver a final reply.
+    let handler: SessionEventHandler | undefined;
+    const session: StubSession = {
+      subscribe: (fn) => {
+        handler = fn;
+        return () => {};
+      },
+    };
+
+    const subscription = subscribeEmbeddedPiSession({
+      session: session as unknown as Parameters<
+        typeof subscribeEmbeddedPiSession
+      >[0]["session"],
+      runId: "run",
+      blockReplyChunking: { minChars: 50, maxChars: 200 }, // Chunking enabled
+    });
+
+    // Simulate non-streaming model: only message_start and message_end, no text_delta
+    handler?.({ type: "message_start", message: { role: "assistant" } });
+
+    const assistantMessage = {
+      role: "assistant",
+      content: [{ type: "text", text: "Response from non-streaming model" }],
+    } as AssistantMessage;
+
+    handler?.({ type: "message_end", message: assistantMessage });
+
+    expect(subscription.assistantTexts).toEqual([
+      "Response from non-streaming model",
+    ]);
+  });
+
   it("does not append when text_end content is a prefix of deltas", () => {
     let handler: ((evt: unknown) => void) | undefined;
     const session: StubSession = {

src/agents/pi-embedded-subscribe.ts

@@ -516,8 +516,10 @@ export function subscribeEmbeddedPiSession(params: {
 
           const addedDuringMessage =
             assistantTexts.length > assistantTextBaseline;
-          const chunkingEnabled = Boolean(blockChunking);
-          if (!chunkingEnabled && !addedDuringMessage && text) {
+          const chunkerHasBuffered = blockChunker?.hasBuffered() ?? false;
+          // Non-streaming models (no text_delta): ensure assistantTexts gets the
+          // final text when the chunker has nothing buffered to drain.
+          if (!addedDuringMessage && !chunkerHasBuffered && text) {
             const last = assistantTexts.at(-1);
             if (!last || last !== text) assistantTexts.push(text);
           }