Skip to content
CC Switch WikiUnofficial Community Wiki

Volcengine Agent Plan config.toml mismatch with official docs - CC Switch

Quick fix

Modify config.toml: set model_provider and section to volcengine-agent-plan; use env_key auth.

Symptom

Section titled “Symptom”
Error output
None

Cause

Section titled “Cause”

The `generateThirdPartyConfig()` generic template in cc-switch generates a `config.toml` for Volcengine Agent Plan that conflicts with the official Volcengine integration docs (updated 2026-07). Key differences: the template uses `model_provider = "custom"` and `[model_providers.custom]`, whereas the official docs require `model_provider = "volcengine-agent-plan"`; the template uses `requires_openai_auth = true` for auth, while the official docs recommend `env_key = "ARK_API_KEY"`; additionally, the template hardcodes `model_reasoning_effort = "high"` and adds `disable_response_storage = true`, which are absent or optional in the official docs. Due to this mismatch, users must manually correct the configuration file to use Codex.

Furthermore, even after manual correction, cc-switch's automatic model list fetching fails. This feature relies on the OpenAI-compatible `{base_url}/models` endpoint, which Volcengine Agent Plan API does not provide (returns 404). Volcengine offers a dedicated `ListArkAgentPlanModel` API, but it uses IAM signature authentication, differing from the Agent Plan's Bearer Token system, preventing direct invocation by cc-switch.

Fix

Section titled “Fix”

  1. Open the `config.toml` file generated by cc-switch and locate the Volcengine Agent Plan configuration section.

    config.toml
    [model_providers.volcengine-agent-plan]
    name = "volcengine-agent-plan"
    base_url = "https://ark.cn-beijing.volces.com/api/plan/v3"
    wire_api = "responses"
    env_key = "ARK_API_KEY"

  2. Ensure the global `model_provider` is set to volcengine-agent-plan and add `model_supports_reasoning_summaries = true`. Remove template defaults like `model_reasoning_effort` and `disable_response_storage = true`.

    config.toml
    model_provider = "volcengine-agent-plan"
    model = "<Model_Name>"
    model_supports_reasoning_summaries = true

Affected Versions

Section titled “Affected Versions”
ToolCodex
Version3.19.2
PlatformsWindows

Source Issues

Section titled “Source Issues”

This page is distilled from 2 real issues

FAQ

Section titled “FAQ”
Why can't cc-switch automatically fetch the model list after I manually corrected config.toml?
Because cc-switch's automatic model list fetching depends on the OpenAI-compatible `/models` endpoint, which Volcengine Agent Plan API does not provide. Volcengine has a dedicated `ListArkAgentPlanModel` API, but the authentication method differs, and cc-switch currently does not support it.
Does Volcengine Coding Plan have the same issue?
Yes, the preset for Volcengine Coding Plan also has similar inconsistencies. The official docs require `model_provider = "volcengine-coding-plan"` and corresponding `env_key`. It is recommended to adjust manually following the Agent Plan correction method.
Section titled “Related problems”
Auth conflictLog out of the official Claude login, clear conflicting ANTHROPIC_* variables, set hasCompletedOnboarding to true in ~/.claude.json, and restart the terminal.6 issues mergedANTHROPIC_AUTH_TOKEN and ANTHROPIC_API_KEYOpen `~/.claude/settings.json`, delete the duplicate auth variable so only the one your provider needs remains, save, and restart Claude Code.4 issues mergedModelScope reports 401 token invalid - Claude CodeAdd ANTHROPIC_MODEL to the provider environment variables in cc-switch to resolve the 401. It is not synced to settings.json by design.2 issues merged

This is an unofficial community wiki with no affiliation to the cc-switch authors or the project itself. Its content is compiled from the project's public GitHub issues. This site distributes no software.