kunthawat/moreminimore-vibe
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

Add LM Studio support (#22)

Browse Source
This commit is contained in:
Piotr Wilkin (ilintar)
2025-05-02 23:51:32 +02:00
committed by GitHub
parent 3529627172
commit 5fc49231ee
15 changed files with 394 additions and 123 deletions
Download Patch File Download Diff File Expand all files Collapse all files

46
package-lock.json generated
View File

@@ -12,6 +12,7 @@
"@ai-sdk/anthropic": "^1.2.8",
"@ai-sdk/google": "^1.2.10",
"@ai-sdk/openai": "^1.3.7",
"@ai-sdk/openai-compatible": "^0.2.13",
"@biomejs/biome": "^1.9.4",
"@dyad-sh/supabase-management-js": "v1.0.0",
"@monaco-editor/react": "^4.7.0-rc.0",
@@ -104,6 +105,51 @@
"node": ">=20"
}
},
"node_modules/@ai-sdk/openai-compatible": {
"version": "0.2.13",
"resolved": "https://registry.npmjs.org/@ai-sdk/openai-compatible/-/openai-compatible-0.2.13.tgz",
"integrity": "sha512-tB+lL8Z3j0qDod/mvxwjrPhbLUHp/aQW+NvMoJaqeTtP+Vmv5qR800pncGczxn5WN0pllQm+7aIRDnm69XeSbg==",
"dev": true,
"dependencies": {
"@ai-sdk/provider": "1.1.3",
"@ai-sdk/provider-utils": "2.2.7"
},
"engines": {
"node": ">=18"
},
"peerDependencies": {
"zod": "^3.0.0"
}
},
"node_modules/@ai-sdk/openai-compatible/node_modules/@ai-sdk/provider": {
"version": "1.1.3",
"resolved": "https://registry.npmjs.org/@ai-sdk/provider/-/provider-1.1.3.tgz",
"integrity": "sha512-qZMxYJ0qqX/RfnuIaab+zp8UAeJn/ygXXAffR5I4N0n1IrvA6qBsjc8hXLmBiMV2zoXlifkacF7sEFnYnjBcqg==",
"dev": true,
"dependencies": {
"json-schema": "^0.4.0"
},
"engines": {
"node": ">=18"
}
},
"node_modules/@ai-sdk/openai-compatible/node_modules/@ai-sdk/provider-utils": {
"version": "2.2.7",
"resolved": "https://registry.npmjs.org/@ai-sdk/provider-utils/-/provider-utils-2.2.7.tgz",
"integrity": "sha512-kM0xS3GWg3aMChh9zfeM+80vEZfXzR3JEUBdycZLtbRZ2TRT8xOj3WodGHPb06sUK5yD7pAXC/P7ctsi2fvUGQ==",
"dev": true,
"dependencies": {
"@ai-sdk/provider": "1.1.3",
"nanoid": "^3.3.8",
"secure-json-parse": "^2.7.0"
},
"engines": {
"node": ">=18"
},
"peerDependencies": {
"zod": "^3.23.8"
}
},
"node_modules/@ai-sdk/anthropic": {
"version": "1.2.8",
"resolved": "https://registry.npmjs.org/@ai-sdk/anthropic/-/anthropic-1.2.8.tgz",

3
package.json
View File

@@ -63,6 +63,7 @@
"vitest": "^3.1.1"
},
"dependencies": {
"@ai-sdk/openai-compatible": "^0.2.13",
"@ai-sdk/anthropic": "^1.2.8",
"@ai-sdk/google": "^1.2.10",
"@ai-sdk/openai": "^1.3.7",
@@ -125,4 +126,4 @@
"update-electron-app": "^3.1.1",
"uuid": "^11.1.0"
}
}
}

4
src/atoms/localModelsAtoms.ts
View File

@@ -4,3 +4,7 @@ import { type LocalModel } from "@/ipc/ipc_types";
export const localModelsAtom = atom<LocalModel[]>([]);
export const localModelsLoadingAtom = atom<boolean>(false);
export const localModelsErrorAtom = atom<Error | null>(null);
export const lmStudioModelsAtom = atom<LocalModel[]>([]);
export const lmStudioModelsLoadingAtom = atom<boolean>(false);
export const lmStudioModelsErrorAtom = atom<Error | null>(null);

176
src/components/ModelPicker.tsx
View File

@@ -19,8 +19,9 @@ import {
import { useEffect, useState } from "react";
import { MODEL_OPTIONS } from "@/constants/models";
import { useLocalModels } from "@/hooks/useLocalModels";
import { useLocalLMSModels } from "@/hooks/useLMStudioModels";
import { ChevronDown } from "lucide-react";
import { LocalModel } from "@/ipc/ipc_types";
interface ModelPickerProps {
selectedModel: LargeLanguageModel;
onModelSelect: (model: LargeLanguageModel) => void;
@@ -31,29 +32,47 @@ export function ModelPicker({
onModelSelect,
}: ModelPickerProps) {
const [open, setOpen] = useState(false);
// Ollama Models Hook
const {
models: localModels,
loading: localModelsLoading,
error: localModelsError,
loadModels,
models: ollamaModels,
loading: ollamaLoading,
error: ollamaError,
loadModels: loadOllamaModels,
} = useLocalModels();
// Load local models when the component mounts or the dropdown opens
// LM Studio Models Hook
const {
models: lmStudioModels,
loading: lmStudioLoading,
error: lmStudioError,
loadModels: loadLMStudioModels,
} = useLocalLMSModels();
// Load models when the dropdown opens
useEffect(() => {
if (open) {
loadModels();
loadOllamaModels();
loadLMStudioModels();
}
}, [open, loadModels]);
}, [open, loadOllamaModels, loadLMStudioModels]);
// Get display name for the selected model
const getModelDisplayName = () => {
if (selectedModel.provider === "ollama") {
return (
localModels.find((model) => model.modelName === selectedModel.name)
ollamaModels.find((model: LocalModel) => model.modelName === selectedModel.name)
?.displayName || selectedModel.name
);
}
if (selectedModel.provider === "lmstudio") {
return (
lmStudioModels.find((model: LocalModel) => model.modelName === selectedModel.name)
?.displayName || selectedModel.name // Fallback to path if not found
);
}
// Fallback for cloud models
return (
MODEL_OPTIONS[selectedModel.provider]?.find(
(model) => model.name === selectedModel.name
@@ -63,8 +82,8 @@ export function ModelPicker({
const modelDisplayName = getModelDisplayName();
// Flatten the model options into a single array with provider information
const allModels = Object.entries(MODEL_OPTIONS).flatMap(
// Flatten the cloud model options
const cloudModels = Object.entries(MODEL_OPTIONS).flatMap(
([provider, models]) =>
models.map((model) => ({
...model,
@@ -72,9 +91,9 @@ export function ModelPicker({
}))
);
// Determine if we have local models available
const hasLocalModels =
!localModelsLoading && !localModelsError && localModels.length > 0;
// Determine availability of local models
const hasOllamaModels = !ollamaLoading && !ollamaError && ollamaModels.length > 0;
const hasLMStudioModels = !lmStudioLoading && !lmStudioError && lmStudioModels.length > 0;
return (
<DropdownMenu open={open} onOpenChange={setOpen}>
@@ -91,12 +110,12 @@ export function ModelPicker({
<ChevronDown className="h-4 w-4" />
</Button>
</DropdownMenuTrigger>
<DropdownMenuContent className="w-56" align="start">
<DropdownMenuContent className="w-64" align="start"> {/* Increased width slightly */}
<DropdownMenuLabel>Cloud Models</DropdownMenuLabel>
<DropdownMenuSeparator />
{/* Cloud models */}
{allModels.map((model) => (
{cloudModels.map((model) => (
<Tooltip key={`${model.provider}-${model.name}`}>
<TooltipTrigger asChild>
<DropdownMenuItem
@@ -135,25 +154,29 @@ export function ModelPicker({
<DropdownMenuSeparator />
{/* Ollama Models Dropdown */}
{/* Ollama Models SubMenu */}
<DropdownMenuSub>
<DropdownMenuSubTrigger
disabled={localModelsLoading || !hasLocalModels}
disabled={ollamaLoading && !hasOllamaModels} // Disable if loading and no models yet
className="w-full font-normal"
>
<div className="flex flex-col items-start">
<span>Local models (Ollama)</span>
{localModelsLoading ? (
{ollamaLoading ? (
<span className="text-xs text-muted-foreground">
Loading...
</span>
) : !hasLocalModels ? (
) : ollamaError ? (
<span className="text-xs text-red-500">
Error loading
</span>
): !hasOllamaModels ? (
<span className="text-xs text-muted-foreground">
None available
</span>
) : (
<span className="text-xs text-muted-foreground">
{localModels.length} models
{ollamaModels.length} models
</span>
)}
</div>
@@ -162,27 +185,32 @@ export function ModelPicker({
<DropdownMenuLabel>Ollama Models</DropdownMenuLabel>
<DropdownMenuSeparator />
{localModelsLoading ? (
{ollamaLoading && ollamaModels.length === 0 ? ( // Show loading only if no models are loaded yet
<div className="text-xs text-center py-2 text-muted-foreground">
Loading models...
</div>
) : localModelsError ? (
<div className="text-xs text-center py-2 text-muted-foreground">
Error loading models
</div>
) : localModels.length === 0 ? (
) : ollamaError ? (
<div className="px-2 py-1.5 text-sm text-red-600">
<div className="flex flex-col">
<span>Error loading models</span>
<span className="text-xs text-muted-foreground">
Is Ollama running?
</span>
</div>
</div>
) : !hasOllamaModels ? (
<div className="px-2 py-1.5 text-sm">
<div className="flex flex-col">
<span>No local models available</span>
<span>No local models found</span>
<span className="text-xs text-muted-foreground">
Start Ollama to use local models
Ensure Ollama is running and models are pulled.
</span>
</div>
</div>
) : (
localModels.map((model) => (
ollamaModels.map((model: LocalModel) => (
<DropdownMenuItem
key={`local-${model.modelName}`}
key={`ollama-${model.modelName}`}
className={
selectedModel.provider === "ollama" &&
selectedModel.name === model.modelName
@@ -208,6 +236,92 @@ export function ModelPicker({
)}
</DropdownMenuSubContent>
</DropdownMenuSub>
{/* LM Studio Models SubMenu */}
<DropdownMenuSub>
<DropdownMenuSubTrigger
disabled={lmStudioLoading && !hasLMStudioModels} // Disable if loading and no models yet
className="w-full font-normal"
>
<div className="flex flex-col items-start">
<span>Local models (LM Studio)</span>
{lmStudioLoading ? (
<span className="text-xs text-muted-foreground">
Loading...
</span>
) : lmStudioError ? (
<span className="text-xs text-red-500">
Error loading
</span>
) : !hasLMStudioModels ? (
<span className="text-xs text-muted-foreground">
None available
</span>
) : (
<span className="text-xs text-muted-foreground">
{lmStudioModels.length} models
</span>
)}
</div>
</DropdownMenuSubTrigger>
<DropdownMenuSubContent className="w-56">
<DropdownMenuLabel>LM Studio Models</DropdownMenuLabel>
<DropdownMenuSeparator />
{lmStudioLoading && lmStudioModels.length === 0 ? ( // Show loading only if no models are loaded yet
<div className="text-xs text-center py-2 text-muted-foreground">
Loading models...
</div>
) : lmStudioError ? (
<div className="px-2 py-1.5 text-sm text-red-600">
<div className="flex flex-col">
<span>Error loading models</span>
<span className="text-xs text-muted-foreground">
{lmStudioError.message} {/* Display specific error */}
</span>
</div>
</div>
) : !hasLMStudioModels ? (
<div className="px-2 py-1.5 text-sm">
<div className="flex flex-col">
<span>No loaded models found</span>
<span className="text-xs text-muted-foreground">
Ensure LM Studio is running and models are loaded.
</span>
</div>
</div>
) : (
lmStudioModels.map((model: LocalModel) => (
<DropdownMenuItem
key={`lmstudio-${model.modelName}`}
className={
selectedModel.provider === "lmstudio" &&
selectedModel.name === model.modelName
? "bg-secondary"
: ""
}
onClick={() => {
onModelSelect({
name: model.modelName,
provider: "lmstudio",
});
setOpen(false);
}}
>
<div className="flex flex-col">
{/* Display the user-friendly name */}
<span>{model.displayName}</span>
{/* Show the path as secondary info */}
<span className="text-xs text-muted-foreground truncate">
{model.modelName}
</span>
</div>
</DropdownMenuItem>
))
)}
</DropdownMenuSubContent>
</DropdownMenuSub>
</DropdownMenuContent>
</DropdownMenu>
);

2
src/constants/models.ts
View File

@@ -8,7 +8,7 @@ export interface ModelOption {
contextWindow?: number;
}
type RegularModelProvider = Exclude<ModelProvider, "ollama">;
type RegularModelProvider = Exclude<ModelProvider, "ollama" | "lmstudio">;
export const MODEL_OPTIONS: Record<RegularModelProvider, ModelOption[]> = {
openai: [
// https://platform.openai.com/docs/models/gpt-4.1

43
src/hooks/useLMStudioModels.ts Normal file
View File

@@ -0,0 +1,43 @@
import { useCallback } from "react";
import { useAtom } from "jotai";
import {
lmStudioModelsAtom,
lmStudioModelsLoadingAtom,
lmStudioModelsErrorAtom,
} from "@/atoms/localModelsAtoms";
import { IpcClient } from "@/ipc/ipc_client";
export function useLocalLMSModels() {
const [models, setModels] = useAtom(lmStudioModelsAtom);
const [loading, setLoading] = useAtom(lmStudioModelsLoadingAtom);
const [error, setError] = useAtom(lmStudioModelsErrorAtom);
const ipcClient = IpcClient.getInstance();
/**
* Load local models from Ollama
*/
const loadModels = useCallback(async () => {
setLoading(true);
try {
const modelList = await ipcClient.listLocalLMStudioModels();
setModels(modelList);
setError(null);
return modelList;
} catch (error) {
console.error("Error loading local LMStudio models:", error);
setError(error instanceof Error ? error : new Error(String(error)));
return [];
} finally {
setLoading(false);
}
}, [ipcClient, setModels, setError, setLoading]);
return {
models,
loading,
error,
loadModels,
};
}

4
src/hooks/useLocalModels.ts
View File

@@ -20,13 +20,13 @@ export function useLocalModels() {
const loadModels = useCallback(async () => {
setLoading(true);
try {
const modelList = await ipcClient.listLocalModels();
const modelList = await ipcClient.listLocalOllamaModels();
setModels(modelList);
setError(null);
return modelList;
} catch (error) {
console.error("Error loading local models:", error);
console.error("Error loading local Ollama models:", error);
setError(error instanceof Error ? error : new Error(String(error)));
return [];
} finally {

81
src/ipc/handlers/local_model_handlers.ts
View File

@@ -1,80 +1,7 @@
import { ipcMain } from "electron";
import log from "electron-log";
import { LocalModelListResponse, LocalModel } from "../ipc_types";
const logger = log.scope("local_model_handlers");
const OLLAMA_API_URL = "http://localhost:11434";
interface OllamaModel {
name: string;
modified_at: string;
size: number;
digest: string;
details: {
format: string;
family: string;
families: string[];
parameter_size: string;
quantization_level: string;
};
}
import { registerOllamaHandlers } from "./local_model_ollama_handler";
import { registerLMStudioHandlers } from "./local_model_lmstudio_handler";
export function registerLocalModelHandlers() {
// Get list of models from Ollama
ipcMain.handle(
"local-models:list",
async (): Promise<LocalModelListResponse> => {
try {
const response = await fetch(`${OLLAMA_API_URL}/api/tags`);
if (!response.ok) {
throw new Error(`Failed to fetch models: ${response.statusText}`);
}
const data = await response.json();
const ollamaModels: OllamaModel[] = data.models || [];
// Transform the data to return just what we need
const models: LocalModel[] = ollamaModels.map((model) => {
// Extract display name by cleaning up the model name
// For names like "llama2:latest" we want to show "Llama 2"
let displayName = model.name.split(":")[0]; // Remove tags like ":latest"
// Capitalize and add spaces for readability
displayName = displayName
.replace(/-/g, " ")
.replace(/(\d+)/, " $1 ") // Add spaces around numbers
.split(" ")
.map((word) => word.charAt(0).toUpperCase() + word.slice(1))
.join(" ")
.trim();
return {
modelName: model.name, // The actual model name used for API calls
displayName, // The user-friendly name
};
});
logger.info(
`Successfully fetched ${models.length} local models from Ollama`
);
return { models, error: null };
} catch (error) {
if (
error instanceof TypeError &&
(error as Error).message.includes("fetch failed")
) {
logger.error("Could not connect to Ollama. Is it running?");
return {
models: [],
error:
"Could not connect to Ollama. Make sure it's running at http://localhost:11434",
};
}
logger.error("Error fetching local models:", error);
return { models: [], error: "Failed to fetch models from Ollama" };
}
}
);
registerOllamaHandlers();
registerLMStudioHandlers();
}

47
src/ipc/handlers/local_model_lmstudio_handler.ts Normal file
View File

@@ -0,0 +1,47 @@
import { ipcMain } from "electron";
import log from "electron-log";
import type { LocalModelListResponse, LocalModel } from "../ipc_types";
const logger = log.scope("lmstudio_handler");
export interface LMStudioModel {
type: "llm" | "embedding" | string;
id: string;
object: string;
publisher: string;
state: "loaded" | "not-loaded";
max_context_length: number;
quantization: string
compatibility_type: string
arch: string;
[key: string]: any;
}
export async function fetchLMStudioModels(): Promise<LocalModelListResponse> {
try {
const modelsResponse: Response = await fetch("http://localhost:1234/api/v0/models");
if (!modelsResponse.ok) {
throw new Error("Failed to fetch models from LM Studio");
}
const modelsJson = await modelsResponse.json();
const downloadedModels = modelsJson.data as LMStudioModel[];
const models: LocalModel[] = downloadedModels
.filter((model: any) => model.type === "llm")
.map((model: any) => ({
modelName: model.id,
displayName: model.id,
provider: "lmstudio"
}));
logger.info(`Successfully fetched ${models.length} models from LM Studio`);
return { models, error: null };
} catch (error) {
return { models: [], error: "Failed to fetch models from LM Studio" };
}
}
export function registerLMStudioHandlers() {
ipcMain.handle('local-models:list-lmstudio', async (): Promise<LocalModelListResponse> => {
return fetchLMStudioModels();
});
}

66
src/ipc/handlers/local_model_ollama_handler.ts Normal file
View File

@@ -0,0 +1,66 @@
import { ipcMain } from "electron";
import log from "electron-log";
import { LocalModelListResponse, LocalModel } from "../ipc_types";
const logger = log.scope("ollama_handler");
const OLLAMA_API_URL = "http://localhost:11434";
interface OllamaModel {
name: string;
modified_at: string;
size: number;
digest: string;
details: {
format: string;
family: string;
families: string[];
parameter_size: string;
quantization_level: string;
};
}
export async function fetchOllamaModels(): Promise<LocalModelListResponse> {
try {
const response = await fetch(`${OLLAMA_API_URL}/api/tags`);
if (!response.ok) {
throw new Error(`Failed to fetch model: ${response.statusText}`);
}
const data = await response.json();
const ollamaModels: OllamaModel[] = data.models || [];
const models: LocalModel[] = ollamaModels.map((model: OllamaModel) => {
const displayName = model.name.split(':')[0]
.replace(/-/g, ' ')
.replace(/(\d+)/, ' $1 ')
.split(' ')
.map(word => word.charAt(0).toUpperCase() + word.slice(1))
.join(' ')
.trim();
return {
modelName: model.name,
displayName,
provider: "ollama",
};
});
logger.info(`Successfully fetched ${models.length} models from Ollama`);
return { models, error: null };
} catch (error) {
if (error instanceof TypeError && (error as Error).message.includes('fetch failed')) {
logger.error("Could not connect to Ollama");
return {
models: [],
error: "Could not connect to Ollama. Make sure it's running at http://localhost:11434"
};
}
return { models: [], error: "Failed to fetch models from Ollama" };
}
}
export function registerOllamaHandlers() {
ipcMain.handle('local-models:list-ollama', async (): Promise<LocalModelListResponse> => {
return fetchOllamaModels();
});
}

28
src/ipc/ipc_client.ts
View File

@@ -785,14 +785,28 @@ export class IpcClient {
}
}
public async listLocalModels(): Promise<LocalModel[]> {
const { models, error } = (await this.ipcRenderer.invoke(
"local-models:list"
)) as LocalModelListResponse;
if (error) {
throw new Error(error);
public async listLocalOllamaModels(): Promise<LocalModel[]> {
try {
const response = await this.ipcRenderer.invoke("local-models:list-ollama");
return response?.models || [];
} catch (error) {
if (error instanceof Error) {
throw new Error(`Failed to fetch Ollama models: ${error.message}`);
}
throw new Error('Failed to fetch Ollama models: Unknown error occurred');
}
}
public async listLocalLMStudioModels(): Promise<LocalModel[]> {
try {
const response = await this.ipcRenderer.invoke("local-models:list-lmstudio");
return response?.models || [];
} catch (error) {
if (error instanceof Error) {
throw new Error(`Failed to fetch LM Studio models: ${error.message}`);
}
throw new Error('Failed to fetch LM Studio models: Unknown error occurred');
}
return models;
}
// Listen for deep link events

1
src/ipc/ipc_types.ts
View File

@@ -94,6 +94,7 @@ export interface SystemDebugInfo {
}
export interface LocalModel {
provider: "ollama" | "lmstudio";
modelName: string; // Name used for API calls (e.g., "llama2:latest")
displayName: string; // User-friendly name (e.g., "Llama 2")
}

12
src/ipc/utils/get_model_client.ts
View File

@@ -3,7 +3,7 @@ import { createGoogleGenerativeAI as createGoogle } from "@ai-sdk/google";
import { createAnthropic } from "@ai-sdk/anthropic";
import { createOpenRouter } from "@openrouter/ai-sdk-provider";
import { createOllama } from "ollama-ai-provider";
import { createOpenAICompatible } from "@ai-sdk/openai-compatible";
import type { LargeLanguageModel, UserSettings } from "../../lib/schemas";
import {
PROVIDER_TO_ENV_VAR,
@@ -82,8 +82,14 @@ export function getModelClient(
case "ollama": {
const provider = createOllama();
return provider(model.name);
}
default: {
}
case "lmstudio": {
// Using LM Studio's OpenAI compatible API
const baseURL = "http://localhost:1234/v1"; // Default LM Studio OpenAI API URL
const provider = createOpenAICompatible({ name: "lmstudio", baseURL });
return provider(model.name);
}
default: {
// Ensure exhaustive check if more providers are added
const _exhaustiveCheck: never = model.provider;
throw new Error(`Unsupported model provider: ${model.provider}`);

1
src/lib/schemas.ts
View File

@@ -36,6 +36,7 @@ export const ModelProviderSchema = z.enum([
"auto",
"openrouter",
"ollama",
"lmstudio",
]);
/**

3
src/preload.ts
View File

@@ -49,7 +49,8 @@ const validInvokeChannels = [
"supabase:list-projects",
"supabase:set-app-project",
"supabase:unset-app-project",
"local-models:list",
"local-models:list-ollama",
"local-models:list-lmstudio",
"window:minimize",
"window:maximize",
"window:close",