v3 M3: z.ai (GLM) provider — Anthropic-compatible poll from a pasted API key

ZaiProvider replaces the zai stub. It reaches z.ai's Anthropic-compatible
endpoint (default https://api.z.ai/api/anthropic) exactly like Claude Code with
ANTHROPIC_BASE_URL pointed at z.ai: a /v1/messages POST with a Bearer key. The
key + base URL come from the config the control panel's z.ai field writes
(providers.zai.base_url/.api_key) — no OAuth, no local files.

Two z.ai-specific cares: (1) static API key, no self-refresh; (2) the GLM Coding
Plan meters by prompts, so a 60s poll would drain quota — the provider
self-throttles to one network call per poll_interval_s (default 15 min) and
serves a cached status in between. Rate-limit parsing assumes z.ai proxies
Anthropic's unified headers (5h->s/sr, 7d->w/wr); if a real response lacks them
it still reports connected and logs the limit-ish headers it did return, so the
mapping can be finalized against a live key.

tools/probe_zai.py dumps a real z.ai response's headers (key via env, never
printed) to pin the exact header names. Graceful without a key: selecting z.ai
shows the blue theme + "no key", never crashes the loop. Registry: all three
providers now real; unknown ids fall back to StubProvider. +7 z.ai tests (mocked
transport: header mapping, auth, throttle, fallback); 120 passed, 2 Linux-only.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

clawdmeter.spec

@@ -41,6 +41,7 @@ hiddenimports = [
     'daemon.providers.base',
     'daemon.providers.anthropic',
     'daemon.providers.openai_codex',
+    'daemon.providers.zai',
     # The exact winrt media modules read_now_playing() pulls in.
     'winrt.windows.media',
     'winrt.windows.media.control',

daemon/providers/__init__.py

@@ -10,12 +10,7 @@ from __future__ import annotations
 from .base import Provider, ProviderStatus, StubProvider
 from .anthropic import AnthropicProvider
 from .openai_codex import OpenAICodexProvider
-
+from .zai import ZaiProvider
-# Brand accents / labels for the not-yet-implemented providers, so the stub still
-# carries the right theme hint to the watch once firmware theming lands.
-_STUB_META = {
-    "zai": ("z.ai GLM", "3859ff"),
-}
 
 
 def get_provider(pid: str) -> Provider:
@@ -23,11 +18,13 @@ def get_provider(pid: str) -> Provider:
         return AnthropicProvider()
     if pid == "openai":
         return OpenAICodexProvider()
-    label, accent = _STUB_META.get(pid, (pid, "d97757"))
+    if pid == "zai":
-    return StubProvider(pid, label=label, accent=accent)
+        return ZaiProvider()
+    # Unknown id (a config from a newer version, say) — a safe placeholder.
+    return StubProvider(pid, label=pid, accent="d97757")
 
 
 __all__ = [
     "Provider", "ProviderStatus", "StubProvider",
-    "AnthropicProvider", "OpenAICodexProvider", "get_provider",
+    "AnthropicProvider", "OpenAICodexProvider", "ZaiProvider", "get_provider",
 ]

daemon/providers/zai.py

@@ -0,0 +1,135 @@
+"""z.ai (GLM Coding Plan) provider (v3 M3).
+
+z.ai exposes an Anthropic-compatible endpoint (default
+``https://api.z.ai/api/anthropic``), so we reach it the same way Claude Code does
+with ``ANTHROPIC_BASE_URL`` pointed at z.ai: a ``/v1/messages`` POST with a Bearer
+API key. The key + base URL come from the config the control panel writes
+(``providers.zai.base_url`` / ``.api_key``) — the "paste your key" field.
+
+Two z.ai-specific cares vs. Anthropic:
+  * Auth is a static API key (Bearer), not the OAuth token — no self-refresh.
+  * The GLM Coding Plan meters by *prompts*, so polling every 60s would burn the
+    user's quota. We self-throttle: hit the network at most once per
+    ``poll_interval_s`` (default 15 min) and serve a cached status in between.
+
+Rate-limit parsing assumes z.ai proxies Anthropic's unified rate-limit headers.
+If a real z.ai response doesn't carry them, poll() still reports "connected" and
+logs the limit-ish headers it DID return, so the mapping can be finalized against
+a live key (see tools/probe_zai.py).
+"""
+
+from __future__ import annotations
+
+import time
+
+import httpx
+
+from .base import Provider, ProviderStatus
+
+DEFAULT_BASE = "https://api.z.ai/api/anthropic"
+DEFAULT_MODEL = "glm-4.6"
+DEFAULT_POLL_INTERVAL_S = 900        # protect the prompt-metered plan: >= 15 min between calls
+
+
+class ZaiProvider(Provider):
+    id = "zai"
+    label = "z.ai GLM"
+    accent = "3859ff"                # brand blue
+
+    def __init__(self, transport: httpx.BaseTransport | None = None) -> None:
+        self._transport = transport   # tests inject an httpx.MockTransport
+        self._cache: ProviderStatus | None = None
+        self._cache_at = 0.0
+
+    async def poll(self) -> ProviderStatus:
+        from daemon.claude_usage_daemon_windows import log
+        try:
+            from daemon.config import load_config, provider_conf
+        except ImportError:
+            from config import load_config, provider_conf
+
+        conf = provider_conf(load_config(), "zai")
+        key = (conf.get("api_key") or "").strip()
+        base = (conf.get("base_url") or "").strip().rstrip("/") or DEFAULT_BASE
+        model = (conf.get("model") or "").strip() or DEFAULT_MODEL
+        try:
+            interval = float(conf.get("poll_interval_s") or DEFAULT_POLL_INTERVAL_S)
+        except (TypeError, ValueError):
+            interval = DEFAULT_POLL_INTERVAL_S
+
+        if not key:
+            return ProviderStatus(ok=False, st="no key", auth_problem=True)
+
+        # Self-throttle: reuse the cached reading until the interval elapses, so the
+        # prompt-metered plan isn't drained by the 60s daemon poll.
+        now = time.time()
+        if self._cache is not None and (now - self._cache_at) < interval:
+            return self._cache
+
+        st = await self._fetch(base, key, model, log)
+        # Cache stable outcomes (fresh data or a definitive auth failure); leave a
+        # transient blip uncached so it retries on the next tick.
+        if st.ok or st.auth_problem:
+            self._cache, self._cache_at = st, now
+        return st
+
+    async def _fetch(self, base: str, key: str, model: str, log) -> ProviderStatus:
+        url = f"{base}/v1/messages"
+        headers = {
+            "Authorization": f"Bearer {key}",
+            "anthropic-version": "2023-06-01",
+            "Content-Type": "application/json",
+        }
+        body = {"model": model, "max_tokens": 1, "messages": [{"role": "user", "content": "hi"}]}
+        client_kw = {"timeout": 20.0}
+        if self._transport is not None:
+            client_kw["transport"] = self._transport
+        try:
+            async with httpx.AsyncClient(**client_kw) as http:
+                resp = await http.post(url, headers=headers, json=body)
+        except httpx.HTTPError as e:
+            log(f"z.ai call failed: {e}")               # transient — no auth flag
+            return ProviderStatus(ok=False, st="offline")
+        if resp.status_code in (401, 403):
+            log(f"z.ai auth rejected: HTTP {resp.status_code}")
+            return ProviderStatus(ok=False, st="bad key", auth_problem=True)
+        if resp.status_code >= 400:
+            log(f"z.ai HTTP {resp.status_code}: {resp.text[:150]}")
+            return ProviderStatus(ok=False, st=f"http {resp.status_code}")
+        return self._parse(resp, log)
+
+    @staticmethod
+    def _parse(resp: httpx.Response, log) -> ProviderStatus:
+        h = resp.headers
+        now = time.time()
+
+        def pct(util: str) -> float:
+            # Anthropic reports utilization as a 0-1 fraction.
+            try:
+                return round(float(util) * 100.0, 1)
+            except (TypeError, ValueError):
+                return 0.0
+
+        def reset_min(ts: str) -> int:
+            try:
+                m = (float(ts) - now) / 60.0
+            except (TypeError, ValueError):
+                return -1
+            return int(round(m)) if m > 0 else -1
+
+        # Assume z.ai proxies Anthropic's unified rate-limit headers.
+        if h.get("anthropic-ratelimit-unified-5h-utilization") is not None:
+            return ProviderStatus(
+                s=pct(h.get("anthropic-ratelimit-unified-5h-utilization")),
+                sr=reset_min(h.get("anthropic-ratelimit-unified-5h-reset")),
+                w=pct(h.get("anthropic-ratelimit-unified-7d-utilization")),
+                wr=reset_min(h.get("anthropic-ratelimit-unified-7d-reset")),
+                st=h.get("anthropic-ratelimit-unified-5h-status", "allowed"),
+                ok=True,
+            )
+        # Connected but no known limit headers — surface what z.ai DID return so the
+        # mapping can be finalized against a live response (tools/probe_zai.py).
+        limitish = sorted(k for k in h.keys()
+                          if any(w in k.lower() for w in ("limit", "rate", "quota", "usage")))
+        log(f"z.ai: no unified rate-limit headers; limit-ish headers = {limitish or '<none>'}")
+        return ProviderStatus(ok=True, st="allowed")

daemon/tests/test_providers.py

@@ -10,7 +10,8 @@ import json
 
 from daemon import config
 from daemon.providers import (
-    get_provider, AnthropicProvider, OpenAICodexProvider, StubProvider, ProviderStatus)
+    get_provider, AnthropicProvider, OpenAICodexProvider, ZaiProvider,
+    StubProvider, ProviderStatus)
 
 
 def _run(coro):
@@ -70,17 +71,19 @@ def test_registry_types():
     a = get_provider("anthropic")
     assert isinstance(a, AnthropicProvider) and a.id == "anthropic"
     o = get_provider("openai")
-    assert isinstance(o, OpenAICodexProvider)
+    assert isinstance(o, OpenAICodexProvider) and o.id == "openai" and o.accent == "10a37f"
-    assert o.id == "openai" and o.accent == "10a37f"
+    z = get_provider("zai")
-    stub = get_provider("zai")          # z.ai is still a stub until M3
+    assert isinstance(z, ZaiProvider) and z.id == "zai" and z.accent == "3859ff"
-    assert isinstance(stub, StubProvider)
+    stub = get_provider("mystery")      # unknown id → safe placeholder, never crashes
-    assert stub.id == "zai" and stub.accent == "3859ff"
+    assert isinstance(stub, StubProvider) and stub.id == "mystery"
 
 
-def test_stub_provider_poll_is_safe():
+def test_zai_poll_without_key_is_safe(tmp_path, monkeypatch):
+    # No api_key configured => "no key", no network call.
+    _write(tmp_path, monkeypatch, {"version": 2, "providers": {"zai": {"enabled": True}}})
     st = _run(get_provider("zai").poll())
     assert isinstance(st, ProviderStatus)
-    assert st.ok is False
+    assert st.ok is False and st.st == "no key"
 
 
 def test_provider_status_payload_keys():

daemon/tests/test_zai.py

@@ -0,0 +1,140 @@
+"""v3 M3 — z.ai (GLM) provider: Anthropic-compatible poll + quota-safe throttle.
+
+Uses httpx.MockTransport so no real z.ai call is made. Covers header mapping,
+auth rejection, the no-known-headers fallback, and the self-throttle cache.
+"""
+
+import asyncio
+import json
+import time
+
+import httpx
+
+from daemon.providers.zai import ZaiProvider, DEFAULT_BASE
+
+
+def _run(coro):
+    # test_zai sorts last; an earlier module's asyncio.run() can leave the shared
+    # loop closed, so fall back to a fresh loop rather than raising (Python 3.13).
+    try:
+        loop = asyncio.get_event_loop()
+        if loop.is_closed():
+            raise RuntimeError
+    except RuntimeError:
+        loop = asyncio.new_event_loop()
+        asyncio.set_event_loop(loop)
+    return loop.run_until_complete(coro)
+
+
+def _config(tmp_path, monkeypatch, zai: dict):
+    p = tmp_path / "config.json"
+    p.write_text(json.dumps({"version": 2, "providers": {"zai": zai}}), encoding="utf-8")
+    monkeypatch.setenv("CLAWDMETER_CONFIG", str(p))
+    monkeypatch.setenv("LOCALAPPDATA", str(tmp_path))
+
+
+def _transport(handler):
+    return httpx.MockTransport(handler)
+
+
+def test_maps_unified_rate_limit_headers(tmp_path, monkeypatch):
+    _config(tmp_path, monkeypatch, {"enabled": True, "api_key": "sk-zai"})
+    now = time.time()
+    seen = {}
+
+    def handler(request: httpx.Request) -> httpx.Response:
+        seen["url"] = str(request.url)
+        seen["auth"] = request.headers.get("authorization")
+        return httpx.Response(200, headers={
+            "anthropic-ratelimit-unified-5h-utilization": "0.42",
+            "anthropic-ratelimit-unified-5h-reset": str(now + 3600),
+            "anthropic-ratelimit-unified-5h-status": "allowed",
+            "anthropic-ratelimit-unified-7d-utilization": "0.10",
+            "anthropic-ratelimit-unified-7d-reset": str(now + 7200),
+        }, json={"ok": True})
+
+    st = _run(ZaiProvider(transport=_transport(handler)).poll())
+    assert st.ok is True
+    assert st.s == 42.0 and 58 <= st.sr <= 60
+    assert st.w == 10.0 and 118 <= st.wr <= 120
+    assert st.st == "allowed"
+    assert seen["url"] == f"{DEFAULT_BASE}/v1/messages"     # default base used
+    assert seen["auth"] == "Bearer sk-zai"                  # Bearer auth
+
+
+def test_no_key_reports_no_key(tmp_path, monkeypatch):
+    _config(tmp_path, monkeypatch, {"enabled": True, "api_key": ""})
+    called = {"n": 0}
+
+    def handler(request):
+        called["n"] += 1
+        return httpx.Response(200)
+
+    st = _run(ZaiProvider(transport=_transport(handler)).poll())
+    assert st.ok is False and st.st == "no key" and st.auth_problem is True
+    assert called["n"] == 0                                 # never hit the network
+
+
+def test_auth_rejection(tmp_path, monkeypatch):
+    _config(tmp_path, monkeypatch, {"enabled": True, "api_key": "bad"})
+
+    def handler(request):
+        return httpx.Response(401, json={"error": "unauthorized"})
+
+    st = _run(ZaiProvider(transport=_transport(handler)).poll())
+    assert st.ok is False and st.st == "bad key" and st.auth_problem is True
+
+
+def test_connected_without_known_headers(tmp_path, monkeypatch):
+    _config(tmp_path, monkeypatch, {"enabled": True, "api_key": "sk"})
+
+    def handler(request):
+        return httpx.Response(200, headers={"x-some-quota": "5"}, json={"ok": True})
+
+    st = _run(ZaiProvider(transport=_transport(handler)).poll())
+    assert st.ok is True and st.st == "allowed" and st.s == 0.0
+
+
+def test_custom_base_url_and_model(tmp_path, monkeypatch):
+    _config(tmp_path, monkeypatch, {
+        "enabled": True, "api_key": "sk", "base_url": "https://zzz.example/anthropic/"})
+    seen = {}
+
+    def handler(request):
+        seen["url"] = str(request.url)
+        seen["body"] = json.loads(request.content)
+        return httpx.Response(200, json={"ok": True})
+
+    _run(ZaiProvider(transport=_transport(handler)).poll())
+    assert seen["url"] == "https://zzz.example/anthropic/v1/messages"   # trailing slash trimmed
+    assert seen["body"]["model"] == "glm-4.6"
+
+
+def test_self_throttle_serves_cache(tmp_path, monkeypatch):
+    _config(tmp_path, monkeypatch, {"enabled": True, "api_key": "sk"})
+    calls = {"n": 0}
+
+    def handler(request):
+        calls["n"] += 1
+        return httpx.Response(200, json={"ok": True})
+
+    p = ZaiProvider(transport=_transport(handler))
+    st1 = _run(p.poll())
+    st2 = _run(p.poll())                                    # within the 15-min interval
+    assert calls["n"] == 1                                 # second served from cache
+    assert st1 is st2
+
+
+def test_transient_error_not_cached(tmp_path, monkeypatch):
+    _config(tmp_path, monkeypatch, {"enabled": True, "api_key": "sk"})
+    calls = {"n": 0}
+
+    def handler(request):
+        calls["n"] += 1
+        raise httpx.ConnectError("boom")
+
+    p = ZaiProvider(transport=_transport(handler))
+    st1 = _run(p.poll())
+    st2 = _run(p.poll())
+    assert st1.ok is False and st1.st == "offline"
+    assert calls["n"] == 2                                 # blip retried, not cached

tools/probe_zai.py

@@ -0,0 +1,54 @@
+#!/usr/bin/env python3
+"""One-off diagnostic: see exactly what z.ai's Anthropic-compatible endpoint
+returns, so the ZaiProvider's rate-limit header mapping can be finalized against
+a real response (the header names aren't documented).
+
+Usage (key via env so it never lands in shell history):
+    ZAI_API_KEY=... python tools/probe_zai.py [base_url] [model]
+
+Defaults: base = https://api.z.ai/api/anthropic , model = glm-4.6
+
+Prints the HTTP status and the FULL response header set (highlighting any header
+mentioning limit/rate/quota/usage/reset), plus a short body preview. Sends a
+1-token "hi" — on a prompt-metered plan that is one prompt, so run it sparingly.
+The API key is read from the environment and never printed.
+"""
+import json
+import os
+import sys
+
+import httpx
+
+BASE = (sys.argv[1] if len(sys.argv) > 1 else "https://api.z.ai/api/anthropic").rstrip("/")
+MODEL = sys.argv[2] if len(sys.argv) > 2 else "glm-4.6"
+KEY = os.environ.get("ZAI_API_KEY") or os.environ.get("ANTHROPIC_AUTH_TOKEN")
+
+if not KEY:
+    sys.exit("Set ZAI_API_KEY in the environment first (it is not printed).")
+
+url = f"{BASE}/v1/messages"
+headers = {
+    "Authorization": f"Bearer {KEY}",
+    "anthropic-version": "2023-06-01",
+    "Content-Type": "application/json",
+}
+body = {"model": MODEL, "max_tokens": 1, "messages": [{"role": "user", "content": "hi"}]}
+
+print(f"POST {url}  (model={MODEL})")
+try:
+    resp = httpx.post(url, headers=headers, json=body, timeout=30.0)
+except httpx.HTTPError as e:
+    sys.exit(f"request failed: {e}")
+
+print(f"HTTP {resp.status_code}\n")
+INTEREST = ("limit", "rate", "quota", "usage", "reset", "remaining")
+print("--- response headers ---")
+for k in sorted(resp.headers.keys()):
+    mark = "  <<<" if any(w in k.lower() for w in INTEREST) else ""
+    print(f"  {k}: {resp.headers[k]}{mark}")
+
+print("\n--- body preview ---")
+try:
+    print(json.dumps(resp.json(), ensure_ascii=False)[:600])
+except ValueError:
+    print(resp.text[:600])