fix(auxiliary): normalize /anthropic base URLs to /v1 for OpenAI-compatible clients by Lempkey · Pull Request #5786 · NousResearch/hermes-agent

Lempkey · 2026-04-07T09:00:33Z

Root cause

MiniMax and MiniMax-CN set inference_base_url to the /anthropic path (e.g. https://api.minimax.io/anthropic) because that is the correct endpoint for the Anthropic Messages API used by the main agent loop.

The auxiliary client constructs an OpenAI() instance using the same URL. The OpenAI SDK then appends /chat/completions, sending every auxiliary/compression request to https://api.minimax.io/anthropic/chat/completions — a 404. The fallback model handles compression instead, at potentially higher cost and lower quality.

What changed

Added _to_openai_base_url() in agent/auxiliary_client.py: strips a terminal /anthropic suffix and replaces it with /v1 before any OpenAI() client is constructed.
Applied at both resolution paths in _resolve_api_key_provider(): the credential pool path (line ~652) and the plain credentials path (line ~669).
No changes to auth.py or ProviderConfig — the normalization lives entirely in the auxiliary client layer where it belongs.

Tests

python -m pytest tests/test_minimax_auxiliary_endpoint.py -v
9 passed

Covers: MiniMax global, MiniMax-CN, trailing slash, non-/anthropic URLs (OpenAI, OpenRouter), /anthropic in domain name (not stripped), /anthropic as mid-path segment (not stripped), empty string, None.

…atible clients MiniMax and MiniMax-CN use an /anthropic path suffix in their inference_base_url (the correct path for the Anthropic Messages API used by the main agent loop). The auxiliary client constructs OpenAI() instances with that same URL; the SDK then appends /chat/completions, producing requests to /anthropic/chat/completions — a 404 on both MiniMax endpoints. Add _to_openai_base_url() which strips a terminal /anthropic suffix and replaces it with /v1 before any OpenAI() client is constructed. Apply it at both resolution paths in _resolve_api_key_provider(): the credential pool path and the plain credentials path. Fixes NousResearch#5781

Salvaged fixes from community PRs: - fix(model_switch): _read_auth_store → _load_auth_store + fix auth store key lookup (was checking top-level dict instead of store['providers']). OAuth providers now correctly detected in /model picker. Cherry-picked from PR #5911 by Xule Lin (linxule). - fix(ollama): pass num_ctx to override 2048 default context window. Ollama defaults to 2048 context regardless of model capabilities. Now auto-detects from /api/show metadata and injects num_ctx into every request. Config override via model.ollama_num_ctx. Fixes #2708. Cherry-picked from PR #5929 by kshitij (kshitijk4poor). - fix(aux): normalize provider aliases for vision/auxiliary routing. Adds _normalize_aux_provider() with 17 aliases (google→gemini, claude→anthropic, glm→zai, etc). Fixes vision routing failure when provider is set to 'google' instead of 'gemini'. Cherry-picked from PR #5793 by e11i (Elizabeth1979). - fix(aux): rewrite MiniMax /anthropic base URLs to /v1 for OpenAI SDK. MiniMax's inference_base_url ends in /anthropic (Anthropic Messages API), but auxiliary client uses OpenAI SDK which appends /chat/completions → 404 at /anthropic/chat/completions. Generic _to_openai_base_url() helper rewrites terminal /anthropic to /v1 for OpenAI-compatible endpoint. Inspired by PR #5786 by Lempkey. Added debug logging to silent exception blocks across all fixes.

…5983) Salvaged fixes from community PRs: - fix(model_switch): _read_auth_store → _load_auth_store + fix auth store key lookup (was checking top-level dict instead of store['providers']). OAuth providers now correctly detected in /model picker. Cherry-picked from PR #5911 by Xule Lin (linxule). - fix(ollama): pass num_ctx to override 2048 default context window. Ollama defaults to 2048 context regardless of model capabilities. Now auto-detects from /api/show metadata and injects num_ctx into every request. Config override via model.ollama_num_ctx. Fixes #2708. Cherry-picked from PR #5929 by kshitij (kshitijk4poor). - fix(aux): normalize provider aliases for vision/auxiliary routing. Adds _normalize_aux_provider() with 17 aliases (google→gemini, claude→anthropic, glm→zai, etc). Fixes vision routing failure when provider is set to 'google' instead of 'gemini'. Cherry-picked from PR #5793 by e11i (Elizabeth1979). - fix(aux): rewrite MiniMax /anthropic base URLs to /v1 for OpenAI SDK. MiniMax's inference_base_url ends in /anthropic (Anthropic Messages API), but auxiliary client uses OpenAI SDK which appends /chat/completions → 404 at /anthropic/chat/completions. Generic _to_openai_base_url() helper rewrites terminal /anthropic to /v1 for OpenAI-compatible endpoint. Inspired by PR #5786 by Lempkey. Added debug logging to silent exception blocks across all fixes. Co-authored-by: Hermes Agent <hermes@nousresearch.com>

teknium1 · 2026-04-08T05:23:51Z

Merged via PR #5983. Your MiniMax URL fix inspiration was salvaged onto current main. Thanks @Lempkey!

…ousResearch#5983) Salvaged fixes from community PRs: - fix(model_switch): _read_auth_store → _load_auth_store + fix auth store key lookup (was checking top-level dict instead of store['providers']). OAuth providers now correctly detected in /model picker. Cherry-picked from PR NousResearch#5911 by Xule Lin (linxule). - fix(ollama): pass num_ctx to override 2048 default context window. Ollama defaults to 2048 context regardless of model capabilities. Now auto-detects from /api/show metadata and injects num_ctx into every request. Config override via model.ollama_num_ctx. Fixes NousResearch#2708. Cherry-picked from PR NousResearch#5929 by kshitij (kshitijk4poor). - fix(aux): normalize provider aliases for vision/auxiliary routing. Adds _normalize_aux_provider() with 17 aliases (google→gemini, claude→anthropic, glm→zai, etc). Fixes vision routing failure when provider is set to 'google' instead of 'gemini'. Cherry-picked from PR NousResearch#5793 by e11i (Elizabeth1979). - fix(aux): rewrite MiniMax /anthropic base URLs to /v1 for OpenAI SDK. MiniMax's inference_base_url ends in /anthropic (Anthropic Messages API), but auxiliary client uses OpenAI SDK which appends /chat/completions → 404 at /anthropic/chat/completions. Generic _to_openai_base_url() helper rewrites terminal /anthropic to /v1 for OpenAI-compatible endpoint. Inspired by PR NousResearch#5786 by Lempkey. Added debug logging to silent exception blocks across all fixes. Co-authored-by: Hermes Agent <hermes@nousresearch.com>

Salvaged fixes from community PRs: - fix(model_switch): _read_auth_store → _load_auth_store + fix auth store key lookup (was checking top-level dict instead of store['providers']). OAuth providers now correctly detected in /model picker. Cherry-picked from PR NousResearch#5911 by Xule Lin (linxule). - fix(ollama): pass num_ctx to override 2048 default context window. Ollama defaults to 2048 context regardless of model capabilities. Now auto-detects from /api/show metadata and injects num_ctx into every request. Config override via model.ollama_num_ctx. Fixes NousResearch#2708. Cherry-picked from PR NousResearch#5929 by kshitij (kshitijk4poor). - fix(aux): normalize provider aliases for vision/auxiliary routing. Adds _normalize_aux_provider() with 17 aliases (google→gemini, claude→anthropic, glm→zai, etc). Fixes vision routing failure when provider is set to 'google' instead of 'gemini'. Cherry-picked from PR NousResearch#5793 by e11i (Elizabeth1979). - fix(aux): rewrite MiniMax /anthropic base URLs to /v1 for OpenAI SDK. MiniMax's inference_base_url ends in /anthropic (Anthropic Messages API), but auxiliary client uses OpenAI SDK which appends /chat/completions → 404 at /anthropic/chat/completions. Generic _to_openai_base_url() helper rewrites terminal /anthropic to /v1 for OpenAI-compatible endpoint. Inspired by PR NousResearch#5786 by Lempkey. Added debug logging to silent exception blocks across all fixes.

…ousResearch#5983) Salvaged fixes from community PRs: - fix(model_switch): _read_auth_store → _load_auth_store + fix auth store key lookup (was checking top-level dict instead of store['providers']). OAuth providers now correctly detected in /model picker. Cherry-picked from PR NousResearch#5911 by Xule Lin (linxule). - fix(ollama): pass num_ctx to override 2048 default context window. Ollama defaults to 2048 context regardless of model capabilities. Now auto-detects from /api/show metadata and injects num_ctx into every request. Config override via model.ollama_num_ctx. Fixes NousResearch#2708. Cherry-picked from PR NousResearch#5929 by kshitij (kshitijk4poor). - fix(aux): normalize provider aliases for vision/auxiliary routing. Adds _normalize_aux_provider() with 17 aliases (google→gemini, claude→anthropic, glm→zai, etc). Fixes vision routing failure when provider is set to 'google' instead of 'gemini'. Cherry-picked from PR NousResearch#5793 by e11i (Elizabeth1979). - fix(aux): rewrite MiniMax /anthropic base URLs to /v1 for OpenAI SDK. MiniMax's inference_base_url ends in /anthropic (Anthropic Messages API), but auxiliary client uses OpenAI SDK which appends /chat/completions → 404 at /anthropic/chat/completions. Generic _to_openai_base_url() helper rewrites terminal /anthropic to /v1 for OpenAI-compatible endpoint. Inspired by PR NousResearch#5786 by Lempkey. Added debug logging to silent exception blocks across all fixes. Co-authored-by: Hermes Agent <hermes@nousresearch.com>

Lempkey mentioned this pull request Apr 7, 2026

[Bug]: MiniMax auxiliary/compression clients route to wrong endpoint (/anthropic instead of /v1) #5781

Closed

1 task

teknium1 mentioned this pull request Apr 8, 2026

fix: provider/model resolution — salvage 4 community PRs + MiniMax aux URL fix #5983

Merged

teknium1 closed this Apr 8, 2026

teknium1 mentioned this pull request Apr 8, 2026

fix: route MiniMax auxiliary/compression clients to /v1 endpoint #5789

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

fix(auxiliary): normalize /anthropic base URLs to /v1 for OpenAI-compatible clients#5786

fix(auxiliary): normalize /anthropic base URLs to /v1 for OpenAI-compatible clients#5786
Lempkey wants to merge 1 commit into
NousResearch:mainfrom
Lempkey:fix/minimax-auxiliary-endpoint-5781

Lempkey commented Apr 7, 2026

Uh oh!

teknium1 commented Apr 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

Lempkey commented Apr 7, 2026

Root cause

What changed

Tests

Uh oh!

teknium1 commented Apr 8, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants