Skip to content
CC Switch WikiUnofficial Community Wiki

No DEEPSEEK settings in the json config file - cc-switch

Quick fix

Upgrade cc-switch to v3.20.0+, tick the reasoning level in the model mapping and save again, to fix config not being written or overwritten by a template.

Symptom

Section titled “Symptom”
Error output
json文件配置里没有相关DEEPSEEK配置项
English translation

There are no DEEPSEEK settings in the json config file

Cause

Section titled “Cause”

When switching or saving a provider in cc-switch (DeepSeek, or a custom openai_responses provider), the config generator has a defect and fails to sync base_url, provider-specific settings, or reasoning_levels into the underlying config file of the target app (Claude, Codex, Hermes). On top of that, when enabling or hot-switching, cc-switch overwrites the target app's catalogue file (~/.codex/cc-switch-model-catalog.json and similar) wholesale from a conservative built-in template — one carrying only none/high reasoning levels, for instance — resetting your custom model parameters and producing model-not-found errors, 429s, or config that simply never applies.

Fix

Section titled “Fix”

  1. Upgrade cc-switch to v3.20.0 or newer to get the fixes for missing config fields and the catalogue overwrite logic.

  2. For a Codex openai_responses custom provider, tick the correct 思考等级 (reasoning level) for each model under 供应商表单 -> 模型映射 (provider form > model mapping) in cc-switch, so it persists into the config store.

  3. Click 启用 (enable) or 保存 (save) again to trigger a config rewrite, so the target app's JSON/YAML config file holds the complete DeepSeek or custom model parameters.

Affected Versions

Section titled “Affected Versions”
Toolcc-switch
Version3.15.0 - 3.19.2
PlatformsWindowsmacOS

Source Issues

Section titled “Source Issues”

This page is distilled from 6 real issues

FAQ

Section titled “FAQ”
Why does ~/.codex/cc-switch-model-catalog.json revert after a restart or switch when I edit it by hand?
cc-switch overwrites that catalogue file wholesale on enable, save, or hot switch to keep state predictable. Do not edit it by hand — tick the reasoning level in the 模型映射 (model mapping) screen in cc-switch v3.20.0+ and let the program write it during generation.
Codex reports 429 too many requests or model not found after switching to DeepSeek. What now?
That usually means older cc-switch did not write base_url or the model mapping correctly, so requests still go to the default endpoint or use the wrong test model. Upgrade to the latest version and save the provider config again to force a refresh of the underlying config file.
Section titled “Related problems”
The Codex Desktop UI project list is completely empty - CodexCC Switch changed the model_provider key name. Set it back to custom, or upgrade. Lost official login and fewer reasoning tiers remain unresolved.4 issues mergedThird-party model still claims based on GPT-5 - CodexMove the `baseInstructions` override out of the `if profile != ProxyChat` block so it applies to ProxyChat. Currently it is not injected even when set.1 issues mergedAPI Error: 400 Param Incorrect Claude CodeUpgrade CC Switch to v3.16.2+ to fix 400 errors caused by missing reasoning_content in MiMo thinking mode.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.