mirrors/hermes-agent
1
0
Fork 0
mirror of https://github.com/NousResearch/hermes-agent.git synced 2026-04-25 00:51:20 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

fix(credential_pool): auto-detect Z.AI endpoint via probe and cache

Browse source 
The credential pool seeder and runtime credential resolver hardcoded
api.z.ai/api/paas/v4 for all Z.AI keys.  Keys on the Coding Plan (or CN
endpoint) would hit the wrong endpoint, causing 401/429 errors on the
first request even though a working endpoint exists.

Add _resolve_zai_base_url() that:
- Respects GLM_BASE_URL env var (no probe when explicitly set)
- Probes all candidate endpoints (global, cn, coding-global, coding-cn)
  via detect_zai_endpoint() to find one that returns HTTP 200
- Caches the detected endpoint in provider state (auth.json) keyed on
  a SHA-256 hash of the API key so subsequent starts skip the probe
- Falls back to the default URL if all probes fail

Wire into both _seed_from_env() in the credential pool and
resolve_api_key_provider_credentials() in the runtime resolver,
matching the pattern from the kimi-coding fix (PR #5566).

Fixes the same class of bug as #5561 but for the zai provider.
This commit is contained in:
Yang Zhi 2026-04-07 00:52:29 +08:00 • committed by Teknium
parent f609bf277d
commit 9e844160f9
3 changed files with 97 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

43
hermes_cli/auth.py
View file

@ -404,6 +404,47 @@ def detect_zai_endpoint(api_key: str, timeout: float = 8.0) -> Optional[Dict[str
return None
def _resolve_zai_base_url(api_key: str, default_url: str, env_override: str) -> str:
"""Return the correct Z.AI base URL by probing endpoints.
If the user has explicitly set GLM_BASE_URL, that always wins.
Otherwise, probe the candidate endpoints to find one that accepts the
key. The detected endpoint is cached in provider state (auth.json) keyed
on a hash of the API key so subsequent starts skip the probe.
"""
if env_override:
return env_override
# Check provider-state cache for a previously-detected endpoint.
auth_store = _load_auth_store()
state = _load_provider_state(auth_store, "zai") or {}
cached = state.get("detected_endpoint")
if isinstance(cached, dict) and cached.get("base_url"):
key_hash = cached.get("key_hash", "")
if key_hash == hashlib.sha256(api_key.encode()).hexdigest()[:16]:
logger.debug("Z.AI: using cached endpoint %s", cached["base_url"])
return cached["base_url"]
# Probe — may take up to ~8s per endpoint.
detected = detect_zai_endpoint(api_key)
if detected and detected.get("base_url"):
# Persist the detection result keyed on the API key hash.
key_hash = hashlib.sha256(api_key.encode()).hexdigest()[:16]
state["detected_endpoint"] = {
"base_url": detected["base_url"],
"endpoint_id": detected.get("id", ""),
"model": detected.get("model", ""),
"label": detected.get("label", ""),
"key_hash": key_hash,
}
_save_provider_state(auth_store, "zai", state)
logger.info("Z.AI: auto-detected endpoint %s (%s)", detected["label"], detected["base_url"])
return detected["base_url"]
logger.debug("Z.AI: probe failed, falling back to default %s", default_url)
return default_url
# =============================================================================
# Error Types
# =============================================================================
@ -2063,6 +2104,8 @@ def resolve_api_key_provider_credentials(provider_id: str) -> Dict[str, Any]:
if provider_id == "kimi-coding":
base_url = _resolve_kimi_base_url(api_key, pconfig.inference_base_url, env_url)
elif provider_id == "zai":
base_url = _resolve_zai_base_url(api_key, pconfig.inference_base_url, env_url)
elif env_url:
base_url = env_url.rstrip("/")
else: