feat: add alibaba-token-plan provider

hermes_cli/auth.py

@@ -347,6 +347,17 @@ class ProviderConfig:
         api_key_env_vars=("ALIBABA_CODING_PLAN_API_KEY", "DASHSCOPE_API_KEY"),
         base_url_env_var="ALIBABA_CODING_PLAN_BASE_URL",
     ),
+    "alibaba-token-plan": ProviderConfig(
+        id="alibaba-token-plan",
+        name="Alibaba Cloud (Token Plan)",
+        auth_type="api_key",
+        # Alibaba's Hermes Agent setup guide recommends the Anthropic-compatible
+        # endpoint for Token Plan. Users can still override this with
+        # ALIBABA_TOKEN_PLAN_BASE_URL if they need the OpenAI-compatible URL.
+        inference_base_url="https://token-plan.ap-southeast-1.maas.aliyuncs.com/apps/anthropic",
+        api_key_env_vars=("ALIBABA_TOKEN_PLAN_API_KEY",),
+        base_url_env_var="ALIBABA_TOKEN_PLAN_BASE_URL",
+    ),
     "minimax-cn": ProviderConfig(
         id="minimax-cn",
         name="MiniMax (China)",
@@ -1428,6 +1439,8 @@ def resolve_provider(
         "minimax-portal": "minimax-oauth", "minimax-global": "minimax-oauth", "minimax_oauth": "minimax-oauth",
         "alibaba_coding": "alibaba-coding-plan", "alibaba-coding": "alibaba-coding-plan",
         "alibaba_coding_plan": "alibaba-coding-plan",
+        "alibaba_token": "alibaba-token-plan", "alibaba-token": "alibaba-token-plan",
+        "alibaba_token_plan": "alibaba-token-plan",
         "claude": "anthropic", "claude-code": "anthropic",
         "github": "copilot", "github-copilot": "copilot",
         "github-models": "copilot", "github-model": "copilot",

hermes_cli/models.py

@@ -415,6 +415,22 @@ def _xai_curated_models() -> list[str]:
         "glm-4.7",
         "MiniMax-M2.5",
     ],
+    # Alibaba Token Plan (Team Edition) — broader model selection including
+    # DeepSeek, latest Qwen, and third-party providers. Singapore region only.
+    # https://modelstudio.console.alibabacloud.com/ap-southeast-1/?tab=doc#/doc/?type=model&url=3028858
+    "alibaba-token-plan": [
+        "qwen3.7-max",
+        "qwen3.6-plus",
+        "qwen3.6-flash",
+        "deepseek-v4-pro",
+        "deepseek-v4-flash",
+        "deepseek-v3.2",
+        "kimi-k2.6",
+        "kimi-k2.5",
+        "glm-5.1",
+        "glm-5",
+        "MiniMax-M2.5",
+    ],
     # Curated HF model list — only agentic models that map to OpenRouter defaults.
     "huggingface": [
         "moonshotai/Kimi-K2.5",
@@ -914,6 +930,7 @@ class ProviderEntry(NamedTuple):
     ProviderEntry("openai-codex",   "OpenAI Codex",             "OpenAI Codex"),
     ProviderEntry("openai-api",     "OpenAI API",               "OpenAI API (api.openai.com, API key)"),
     ProviderEntry("alibaba",        "Qwen Cloud",               "Qwen Cloud / DashScope Coding (Qwen + multi-provider)"),
+    ProviderEntry("alibaba-token-plan", "Alibaba Token Plan",    "Alibaba Cloud Token Plan (Team Edition — Singapore, broad model catalog)"),
     ProviderEntry("xai-oauth",      "xAI Grok OAuth (SuperGrok / Premium+)", "xAI Grok OAuth (SuperGrok / Premium+)"),
     ProviderEntry("xiaomi",         "Xiaomi MiMo",              "Xiaomi MiMo (MiMo-V2.5 and V2 models — pro, omni, flash)"),
     ProviderEntry("tencent-tokenhub", "Tencent TokenHub",       "Tencent TokenHub (Hy3 Preview — direct API via tokenhub.tencentmaas.com)"),
@@ -1010,6 +1027,9 @@ class ProviderEntry(NamedTuple):
     "aliyun": "alibaba",
     "qwen": "alibaba",
     "alibaba-cloud": "alibaba",
+    "alibaba_token": "alibaba-token-plan",
+    "alibaba-token": "alibaba-token-plan",
+    "alibaba_token_plan": "alibaba-token-plan",
     "qwen-portal": "qwen-oauth",
     "gemini-cli": "google-gemini-cli",
     "gemini-oauth": "google-gemini-cli",

hermes_cli/providers.py

@@ -143,6 +143,16 @@ class HermesOverlay:
         transport="openai_chat",
         base_url_env_var="ALIBABA_CODING_PLAN_BASE_URL",
     ),
+    "alibaba-token-plan": HermesOverlay(
+        transport="anthropic_messages",
+        extra_env_vars=("ALIBABA_TOKEN_PLAN_API_KEY",),
+        base_url_override="https://token-plan.ap-southeast-1.maas.aliyuncs.com/apps/anthropic",
+        base_url_env_var="ALIBABA_TOKEN_PLAN_BASE_URL",
+    ),
+    "vercel": HermesOverlay(
+        transport="openai_chat",
+        is_aggregator=True,
+    ),
     "opencode": HermesOverlay(
         transport="openai_chat",
         is_aggregator=True,
@@ -310,6 +320,9 @@ class ProviderDef:
     "alibaba_coding": "alibaba-coding-plan",
     "alibaba-coding": "alibaba-coding-plan",
     "alibaba_coding_plan": "alibaba-coding-plan",
+    "alibaba_token": "alibaba-token-plan",
+    "alibaba-token": "alibaba-token-plan",
+    "alibaba_token_plan": "alibaba-token-plan",
 
     # google-gemini-cli (OAuth + Code Assist)
     "gemini-cli": "google-gemini-cli",
@@ -376,6 +389,7 @@ class ProviderDef:
     "lmstudio": "LM Studio",
     "local": "Local endpoint",
     "bedrock": "AWS Bedrock",
+    "alibaba-token-plan": "Alibaba Token Plan",
     "ollama-cloud": "Ollama Cloud",
     "xai-oauth": "xAI Grok OAuth (SuperGrok / Premium+)",
 }

tests/hermes_cli/test_alibaba_token_plan_provider.py

@@ -0,0 +1,45 @@
+"""Regression tests for Alibaba Token Plan provider wiring."""
+
+from hermes_cli.auth import PROVIDER_REGISTRY, resolve_provider
+from hermes_cli.providers import HERMES_OVERLAYS, TRANSPORT_TO_API_MODE, get_label, get_provider
+from hermes_cli.runtime_provider import _detect_api_mode_for_url
+
+
+TOKEN_PLAN_ANTHROPIC_URL = "https://token-plan.ap-southeast-1.maas.aliyuncs.com/apps/anthropic"
+
+
+def test_auth_registry_uses_token_plan_anthropic_endpoint():
+    cfg = PROVIDER_REGISTRY["alibaba-token-plan"]
+
+    assert cfg.inference_base_url == TOKEN_PLAN_ANTHROPIC_URL
+    assert cfg.api_key_env_vars == ("ALIBABA_TOKEN_PLAN_API_KEY",)
+    assert cfg.base_url_env_var == "ALIBABA_TOKEN_PLAN_BASE_URL"
+
+
+def test_provider_overlay_uses_anthropic_messages_transport():
+    overlay = HERMES_OVERLAYS["alibaba-token-plan"]
+
+    assert overlay.transport == "anthropic_messages"
+    assert TRANSPORT_TO_API_MODE[overlay.transport] == "anthropic_messages"
+    assert overlay.base_url_override == TOKEN_PLAN_ANTHROPIC_URL
+    assert overlay.extra_env_vars == ("ALIBABA_TOKEN_PLAN_API_KEY",)
+
+
+def test_provider_definition_exposes_base_url_key_and_label():
+    provider = get_provider("alibaba_token_plan")
+
+    assert provider is not None
+    assert provider.id == "alibaba-token-plan"
+    assert provider.base_url == TOKEN_PLAN_ANTHROPIC_URL
+    assert provider.api_key_env_vars == ("ALIBABA_TOKEN_PLAN_API_KEY",)
+    assert get_label("alibaba_token") == "Alibaba Token Plan"
+
+
+def test_token_plan_aliases_resolve_to_canonical_provider():
+    assert resolve_provider("alibaba_token") == "alibaba-token-plan"
+    assert resolve_provider("alibaba-token") == "alibaba-token-plan"
+    assert resolve_provider("alibaba_token_plan") == "alibaba-token-plan"
+
+
+def test_token_plan_anthropic_url_auto_detects_api_mode():
+    assert _detect_api_mode_for_url(TOKEN_PLAN_ANTHROPIC_URL) == "anthropic_messages"

tests/hermes_cli/test_model_validation.py

@@ -70,6 +70,12 @@ def test_stepfun_alias_resolved(self):
         assert provider == "stepfun"
         assert model == "step-3.5-flash"
 
+    def test_alibaba_token_plan_alias_resolved(self):
+        provider, model = parse_model_input("alibaba-token:qwen3.7-max", "openrouter")
+        assert provider == "alibaba-token-plan"
+        assert model == "qwen3.7-max"
+        assert provider_label(provider) == "Alibaba Token Plan"
+
     def test_no_slash_no_colon_keeps_provider(self):
         provider, model = parse_model_input("gpt-5.4", "openrouter")
         assert provider == "openrouter"
@@ -606,6 +612,12 @@ def test_zai_known_model_accepted_via_catalog_when_api_down(self):
         assert result["persist"] is True
         assert result["recognized"] is True
 
+    def test_alibaba_token_plan_known_models_accepted_via_catalog_when_api_down(self):
+        result = _validate("deepseek-v4-pro", provider="alibaba-token-plan", api_models=None)
+        assert result["accepted"] is True
+        assert result["persist"] is True
+        assert result["recognized"] is True
+
     def test_unknown_provider_soft_accepted_when_api_down(self):
         # No catalog for unknown providers — soft-accept with a Note.
         with patch("hermes_cli.models.provider_model_ids", return_value=[]):