使提供者和模型的配置更加通用

目前的配置屏幕和文件迫使您选择一个具有多个模型字符串的单一的ollama url,或者三个前沿提供者之一,而不存在明确的模型配置的可能性.

作为AI工程师,我使用许多供应商,包括ollama,LM Studio,以及几个公司推论服务器.

Pi编码套有优雅的格式,可以用作灵感:

{
"提供商":{
"lmstudio":{{.
" BaseUrl":"http://localhost:1234/v1","http","http".
"Api":"OpenAI - 补齐",
"阿皮凯":"sk live xxx","sk".
"相会":{
"支持开发者Role":虚假,
"支持Reasoning Effort":虚假
{\fn华文楷体\fs16\1cHE0E0E0}
""模型": [
{
"id":"google/gemma-4-e4b","go".
"推理":真实
{\fn华文楷体\fs16\1cHE0E0E0}
{
"id":"google/gemma-4-31b-qat","go".
"推理":真实
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?
[ . ]
{\fn华文楷体\fs16\1cHE0E0E0}
"军"一:{
" BaseUrl":"https://inference.corphost.dk/v1", (中文(简体) ).
"api":"OpenAI-完成",".
"ApiKey":"sk live yyy","sk","sk".
"相会":{
"支持开发者Role":虚假,
"支持Reasoning Effort":虚假
{\fn华文楷体\fs16\1cHE0E0E0}
""模型": [
{
"id":"qwen3.5-397b","q".
"外出":{
"聊天 template kwargs":{
"能 想":虚假
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?
[ . ]
{\fn华文楷体\fs16\1cHE0E0E0}
"军":{
" BaseUrl":"https://inference2.corpshost.dk/v1", (中文(简体) ).
"api":"OpenAI-完成",".
"apiKey":"sk live zzz","sk".
"相会":{
"支持开发者Role":虚假,
"支持Reasoning Effort":虚假
{\fn华文楷体\fs16\1cHE0E0E0}
""模型": [
{
"id":"Qwen-235b","ZZ".
"外出":{
"聊天 template kwargs":{
"能 想":虚假
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?
[ . ]
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?
{\fn黑体\fs22\bord1\shad0\3aHBE\4aH00\fscx67\fscy66\2cHFFFFFF\3cH808080}你觉得呢?