APEX/BrowserUse_and_ComputerUse_skills
5
0
Fork 0
Code Issues 6 Pull requests 1 Projects 1 Releases Packages Wiki Activity Actions

feat: allow custom endpoints to use responses API via api_mode override (#1651)

Browse source 
Add HERMES_API_MODE env var and model.api_mode config field to let
custom OpenAI-compatible endpoints opt into codex_responses mode
without requiring the OpenAI Codex OAuth provider path.

- _get_configured_api_mode() reads HERMES_API_MODE env (precedence)
  then model.api_mode from config.yaml; validates against whitelist
- Applied in both _resolve_openrouter_runtime() and
  _resolve_named_custom_runtime() (original PR only covered openrouter)
- Fix _dump_api_request_debug() to show /responses URL when in
  codex_responses mode instead of always showing /chat/completions
- Tests for config override, env override, invalid values, named
  custom providers, and debug dump URL for both API modes

Inspired by PR #1041 by @mxyhi.

Co-authored-by: mxyhi <mxyhi@users.noreply.github.com>
This commit is contained in:
Teknium 2026-03-17 02:04:36 -07:00 • committed by GitHub
parent 68fbcdaa06
commit f2414bfd45
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
4 changed files with 131 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

71
tests/test_runtime_provider_resolution.py
View file

@ -326,3 +326,74 @@ def test_resolve_requested_provider_precedence(monkeypatch):
monkeypatch.delenv("HERMES_INFERENCE_PROVIDER", raising=False)
assert rp.resolve_requested_provider() == "auto"
# ── api_mode override tests ─────────────────────────────────────────────
def test_custom_endpoint_api_mode_from_config(monkeypatch):
"""model.api_mode in config.yaml should override the default chat_completions."""
monkeypatch.setattr(rp, "resolve_provider", lambda *a, **k: "openrouter")
monkeypatch.setattr(
rp, "_get_model_config",
lambda: {
"provider": "custom",
"base_url": "http://127.0.0.1:9208/v1",
"api_mode": "codex_responses",
},
)
monkeypatch.setenv("OPENAI_BASE_URL", "http://127.0.0.1:9208/v1")
monkeypatch.setenv("OPENAI_API_KEY", "test-key")
monkeypatch.delenv("OPENROUTER_BASE_URL", raising=False)
monkeypatch.delenv("OPENROUTER_API_KEY", raising=False)
monkeypatch.delenv("HERMES_API_MODE", raising=False)
resolved = rp.resolve_runtime_provider(requested="custom")
assert resolved["api_mode"] == "codex_responses"
assert resolved["base_url"] == "http://127.0.0.1:9208/v1"
def test_env_api_mode_overrides_config(monkeypatch):
"""HERMES_API_MODE env var takes precedence over config."""
monkeypatch.setattr(rp, "resolve_provider", lambda *a, **k: "openrouter")
monkeypatch.setattr(rp, "_get_model_config", lambda: {"api_mode": "chat_completions"})
monkeypatch.setenv("OPENAI_BASE_URL", "http://127.0.0.1:9208/v1")
monkeypatch.setenv("OPENAI_API_KEY", "test-key")
monkeypatch.setenv("HERMES_API_MODE", "codex_responses")
monkeypatch.delenv("OPENROUTER_BASE_URL", raising=False)
monkeypatch.delenv("OPENROUTER_API_KEY", raising=False)
resolved = rp.resolve_runtime_provider(requested="custom")
assert resolved["api_mode"] == "codex_responses"
def test_invalid_api_mode_ignored(monkeypatch):
"""Invalid api_mode values should fall back to chat_completions."""
monkeypatch.setattr(rp, "resolve_provider", lambda *a, **k: "openrouter")
monkeypatch.setattr(rp, "_get_model_config", lambda: {"api_mode": "bogus_mode"})
monkeypatch.setenv("OPENAI_BASE_URL", "http://127.0.0.1:9208/v1")
monkeypatch.setenv("OPENAI_API_KEY", "test-key")
monkeypatch.delenv("HERMES_API_MODE", raising=False)
monkeypatch.delenv("OPENROUTER_BASE_URL", raising=False)
monkeypatch.delenv("OPENROUTER_API_KEY", raising=False)
resolved = rp.resolve_runtime_provider(requested="custom")
assert resolved["api_mode"] == "chat_completions"
def test_named_custom_provider_respects_api_mode(monkeypatch):
"""Named custom providers should also pick up api_mode overrides."""
monkeypatch.setattr(rp, "resolve_provider", lambda *a, **k: "my-server")
monkeypatch.setattr(
rp, "_get_named_custom_provider",
lambda p: {"name": "my-server", "base_url": "http://localhost:8000/v1", "api_key": "sk-test"},
)
monkeypatch.setenv("HERMES_API_MODE", "codex_responses")
resolved = rp.resolve_runtime_provider(requested="my-server")
assert resolved["api_mode"] == "codex_responses"
assert resolved["base_url"] == "http://localhost:8000/v1"