classifyTopic fails when summarizer uses Anthropic endpoint (Kimi Code)

[Lunareith]

Bug Report: classifyTopic fails when summarizer is configured with Anthropic endpoint

Environment

• OpenClaw version: 2026.4.29
• MemOS plugin: memos-local-openclaw-plugin v1.0.8
• Model: anthropic/kimi-for-coding
• Endpoint: https://api.kimi.com/coding/v1/messages

Problem

TaskProcessor.classifyTopic always fails with 404 when the MemOS summarizer is configured to use an Anthropic-compatible endpoint (e.g., Kimi Code's /messages endpoint).

Error Log

classifyTopic failed (anthropic/kimi-for-coding), trying next:
Error: OpenAI topic-classifier failed (404):
{"error":{"message":"The requested resource was not found","type":"resource_not_found_error"}}

Root Cause

classifyTopic in src/ingest/providers/index.ts always calls classifyTopicOpenAI(), which sends requests in OpenAI format (/chat/completions), regardless of the summarizer's configured provider.

Component	Format Used	Endpoint
summarizer (configured)	Anthropic Messages	/coding/v1/messages ✅
classifyTopic (hardcoded)	OpenAI Chat Completions	/coding/v1 ❌

When the endpoint only supports Anthropic format, the OpenAI-style classifyTopic request returns 404.

Why This Didn't Happen Before

• DeepSeek supports both OpenAI and Anthropic formats simultaneously, so OpenAI-style classifyTopic works even when summarizer uses Anthropic config.
• Kimi Code's Anthropic endpoint (/coding/v1/messages) only accepts Anthropic format, rejecting OpenAI-style requests.

Impact

• Every incoming message triggers a classifyTopic call that fails and falls back.
• Wastes tokens (small per-call, but adds up).
• Pollutes logs with repeated 404 errors.
• Increases event-loop latency (related to observed Gateway slowdowns).

Proposed Fix

Option A: Make classifyTopic respect the summarizer's configured provider and call the appropriate implementation (OpenAI vs Anthropic).

Option B: Add a separate topicClassifier configuration key so users can explicitly set a different provider/model for topic classification.

Related Code

• src/ingest/providers/index.ts: classifyTopic()
• src/ingest/providers/openai.ts: classifyTopicOpenAI()
• src/ingest/task-processor.ts: onChunksIngested() → classifyTopic()

---

Additional context: This was discovered while investigating why QQ Bot responses became slow. The repeated classifyTopic failures contribute to event-loop blocking in the OpenClaw Gateway.

[Memtensor-AI]

🤖 AutoDev has picked up this issue and started working on it.

Task ID: f36660c380558d6e
Working branch: bugfix/autodev-1611
Target branch: latest dev* branch
Workflow: opsp (analysis → coding → testing → PR)

I will post the PR link here once done. If I need more information, I will ask in the comments.

[Memtensor-AI]

✅ AutoDev task f36660c380558d6e completed.

Summary: Fixes #1611 — Summarizer.classifyTopic and Summarizer.arbitrateTopicSplit previously dispatched every call to the OpenAI helper, which produced a 404 whenever the summarizer was wired to an Anthropic-only endpoint such as Kimi Code's /coding/v1/messages. Each ingested message triggered the failure, wasting tokens, polluting logs, and contributing to event-loop slowdowns in the OpenClaw Gateway.

The fix mirrors the existing per-provider pattern already used for summarize / judgeNewTopic / filterRelevant / judgeDedup. Six new helpers (classifyTopic{Anthropic,Gemini,Bedrock} + arbitrateTopicSplit{Anthropic,Gemini,Bedrock}) are added to the matching provider modules; TOPIC_CLASSIFIER_PROMPT and TOPIC_ARBITRATION_PROMPT are exported from openai.ts so all four providers share one canonical prompt; and the two switch statements in apps/memos-local-openclaw/src/ingest/providers/index.ts now route by cfg.provider. The byte-identical copy in packages/memos-core/src/ingest/providers/ is updated in lockstep so the bug cannot reappear after the next sync.

Verification: new dispatch test apps/memos-local-openclaw/tests/topic-classifier-dispatch.test.ts (8 cases, all green) stubs globalThis.fetch and asserts URL + body shape per provider — including a regression case that proves an Anthropic 404 now surfaces an Anthropic topic-classifier failed error rather than the old OpenAI topic-classifier failed. Full vitest suite: 230/235 pass (the 5 failures are pre-existing on the baseline branch and unrelated to topic classification — accuracy benchmark, skill auto-install default, sqlite test-data ordering, and update-install signal handling). tsc --noEmit is clean.

Base branch: main
Branch: bugfix/autodev-1611
Commit: 16e9e4b
PR: #1997
Assigned to: @MatthewZhuang, @CarltonXiang, @syzsunshine219, @World-controller
Reviewers: @MatthewZhuang, @CarltonXiang, @syzsunshine219, @World-controller