nex-code 0.3.0 → 0.3.2

package/README.md

@@ -13,6 +13,7 @@
 </p>
 
 <p align="center">
+  <a href="https://www.npmjs.com/package/nex-code"><img src="https://img.shields.io/npm/v/nex-code.svg" alt="npm version"></a>
   <a href="https://github.com/hybridpicker/nex-code/actions/workflows/ci.yml"><img src="https://github.com/hybridpicker/nex-code/actions/workflows/ci.yml/badge.svg" alt="CI"></a>
   <a href="https://opensource.org/licenses/MIT"><img src="https://img.shields.io/badge/License-MIT-blue.svg" alt="License: MIT"></a>
   <img src="https://img.shields.io/badge/node-%3E%3D18-brightgreen.svg" alt="Node >= 18">
@@ -25,10 +26,15 @@
 ## Quickstart
 
 ```bash
-git clone https://github.com/hybridpicker/nex-code.git
-cd nex-code && npm install
-cp .env.example .env       # add at least one API key
-npm start                   # or: npm link && nex-code
+npx nex-code
+```
+
+Or install globally:
+
+```bash
+npm install -g nex-code
+cd ~/your-project
+nex-code
 ```
 
 That's it. You'll see the banner, your project context, and the `>` prompt. Start typing.
@@ -55,10 +61,22 @@ That's it. You'll see the banner, your project context, and the `>` prompt. Star
 - Node.js 18+
 - At least one API key **or** a local [Ollama](https://ollama.com/download) server
 
-### Installation
+### Install from npm
+
+```bash
+npm install -g nex-code
+```
+
+Or run directly without installing:
 
 ```bash
-git clone git@github.com:hybridpicker/nex-code.git
+npx nex-code
+```
+
+### Install from source (for contributors)
+
+```bash
+git clone https://github.com/hybridpicker/nex-code.git
 cd nex-code
 npm install
 cp .env.example .env
@@ -68,7 +86,7 @@ npm run install-hooks
 
 ### Configure a Provider
 
-Add one or more API keys to `.env`:
+Create a `.env` file in your project directory (or set environment variables):
 
 ```bash
 # Pick any — only one is required
@@ -682,6 +700,13 @@ CI runs on GitHub Actions (Node 18/20/22).
 
 Everything else is Node.js built-in.
 
+## Installation
+
+```bash
+npm install -g nex-code    # global install
+npx nex-code               # or run without installing
+```
+
 ## License
 
 MIT

package/cli/costs.js

@@ -33,12 +33,17 @@ const PRICING = {
     'gemini-2.0-flash-lite': { input: 0.075, output: 0.30 },
   },
   'ollama': {
-    'kimi-k2.5':      { input: 0, output: 0 },
-    'qwen3-coder':    { input: 0, output: 0 },
-    'deepseek-r1':    { input: 0, output: 0 },
-    'llama4':         { input: 0, output: 0 },
-    'qwen3:30b-a3b':  { input: 0, output: 0 },
-    'devstral':       { input: 0, output: 0 },
+    'qwen3-coder:480b':     { input: 0, output: 0 },
+    'qwen3-coder-next':     { input: 0, output: 0 },
+    'devstral-2:123b':      { input: 0, output: 0 },
+    'devstral-small-2:24b': { input: 0, output: 0 },
+    'kimi-k2.5':            { input: 0, output: 0 },
+    'kimi-k2:1t':           { input: 0, output: 0 },
+    'deepseek-v3.2':        { input: 0, output: 0 },
+    'minimax-m2.5':         { input: 0, output: 0 },
+    'glm-5':                { input: 0, output: 0 },
+    'glm-4.7':              { input: 0, output: 0 },
+    'gpt-oss:120b':         { input: 0, output: 0 },
   },
   'local': {},
 };

package/cli/index.js

@@ -914,18 +914,26 @@ function startREPL() {
         const { execSync } = require('child_process');
         execSync('curl -s --max-time 2 http://localhost:11434/api/tags', { encoding: 'utf-8', stdio: 'pipe' });
         setActiveModel('local:llama3');
-        console.log(`${C.green}Local Ollama detected — using local models${C.reset}`);
-        console.log(`${C.dim}Set API keys for cloud providers: OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY${C.reset}\n`);
+        console.log(`${C.green}✓ Local Ollama detected — using local models${C.reset}`);
+        console.log(`${C.dim}Tip: Set API keys for cloud providers for more model options (OPENAI_API_KEY, ANTHROPIC_API_KEY, etc.)${C.reset}\n`);
         localDetected = true;
       } catch {
         // Local Ollama not available
       }
     }
     if (!localDetected) {
-      console.error(`${C.red}No provider configured and no local Ollama running.${C.reset}`);
-      console.error(`${C.gray}Options:${C.reset}`);
-      console.error(`${C.gray}  1. Install Ollama: https://ollama.com/download${C.reset}`);
-      console.error(`${C.gray}  2. Set an API key: OPENAI_API_KEY, ANTHROPIC_API_KEY, GEMINI_API_KEY, or OLLAMA_API_KEY${C.reset}`);
+      console.error(`\n${C.red}✗ No provider configured and no local Ollama detected.${C.reset}\n`);
+      console.error(`${C.white}nex-code needs at least one LLM provider to work.${C.reset}\n`);
+      console.error(`${C.white}Option 1 — Free local models (no API key needed):${C.reset}`);
+      console.error(`${C.gray}  Install Ollama:  ${C.cyan}https://ollama.com/download${C.reset}`);
+      console.error(`${C.gray}  Pull a model:    ${C.cyan}ollama pull qwen3-coder${C.reset}`);
+      console.error(`${C.gray}  Then restart:    ${C.cyan}nex-code${C.reset}\n`);
+      console.error(`${C.white}Option 2 — Cloud providers (set one in .env or as env var):${C.reset}`);
+      console.error(`${C.gray}  OLLAMA_API_KEY=...     ${C.dim}# Ollama Cloud (free tier available)${C.reset}`);
+      console.error(`${C.gray}  OPENAI_API_KEY=...     ${C.dim}# OpenAI${C.reset}`);
+      console.error(`${C.gray}  ANTHROPIC_API_KEY=...  ${C.dim}# Anthropic${C.reset}`);
+      console.error(`${C.gray}  GEMINI_API_KEY=...     ${C.dim}# Google Gemini${C.reset}\n`);
+      console.error(`${C.dim}Create a .env file in your project directory or export the variable.${C.reset}`);
       process.exit(1);
     }
   }

package/cli/ollama.js

@@ -9,7 +9,7 @@ const registry = require('./providers/registry');
 
 const MODELS = {
   'kimi-k2.5': { id: 'kimi-k2.5', name: 'Kimi K2.5', max_tokens: 16384 },
-  'qwen3-coder': { id: 'qwen3-coder', name: 'Qwen3 Coder', max_tokens: 16384 },
+  'qwen3-coder:480b': { id: 'qwen3-coder:480b', name: 'Qwen3 Coder 480B', max_tokens: 16384 },
 };
 
 function getActiveModel() {

package/cli/providers/ollama.js

@@ -7,20 +7,34 @@ const axios = require('axios');
 const { BaseProvider } = require('./base');
 
 const OLLAMA_MODELS = {
-  // Primary: Best coding model for agentic workflows
-  'qwen3-coder': { id: 'qwen3-coder', name: 'Qwen3 Coder', maxTokens: 16384, contextWindow: 131072 },
+  // Primary: Best coding models for agentic workflows
+  'qwen3-coder:480b': { id: 'qwen3-coder:480b', name: 'Qwen3 Coder 480B', maxTokens: 16384, contextWindow: 131072 },
   'qwen3-coder-next': { id: 'qwen3-coder-next', name: 'Qwen3 Coder Next', maxTokens: 16384, contextWindow: 131072 },
-  // Reasoning specialists
-  'deepseek-r1': { id: 'deepseek-r1', name: 'DeepSeek R1', maxTokens: 16384, contextWindow: 131072 },
-  'deepseek-r1:14b': { id: 'deepseek-r1:14b', name: 'DeepSeek R1 14B', maxTokens: 8192, contextWindow: 128000 },
-  // Agent-focused models
-  'devstral': { id: 'devstral', name: 'Devstral', maxTokens: 16384, contextWindow: 131072 },
+  'devstral-2:123b': { id: 'devstral-2:123b', name: 'Devstral 2 123B', maxTokens: 16384, contextWindow: 131072 },
+  'devstral-small-2:24b': { id: 'devstral-small-2:24b', name: 'Devstral Small 2 24B', maxTokens: 16384, contextWindow: 131072 },
+  // Large general-purpose models
+  'kimi-k2.5': { id: 'kimi-k2.5', name: 'Kimi K2.5', maxTokens: 16384, contextWindow: 256000 },
+  'kimi-k2:1t': { id: 'kimi-k2:1t', name: 'Kimi K2 1T', maxTokens: 16384, contextWindow: 256000 },
+  'kimi-k2-thinking': { id: 'kimi-k2-thinking', name: 'Kimi K2 Thinking', maxTokens: 16384, contextWindow: 256000 },
+  'deepseek-v3.2': { id: 'deepseek-v3.2', name: 'DeepSeek V3.2', maxTokens: 16384, contextWindow: 131072 },
+  'deepseek-v3.1:671b': { id: 'deepseek-v3.1:671b', name: 'DeepSeek V3.1 671B', maxTokens: 16384, contextWindow: 131072 },
+  'cogito-2.1:671b': { id: 'cogito-2.1:671b', name: 'Cogito 2.1 671B', maxTokens: 16384, contextWindow: 131072 },
+  // Medium models
+  'qwen3-next:80b': { id: 'qwen3-next:80b', name: 'Qwen3 Next 80B', maxTokens: 16384, contextWindow: 131072 },
+  'qwen3.5:397b': { id: 'qwen3.5:397b', name: 'Qwen3.5 397B', maxTokens: 16384, contextWindow: 131072 },
+  'mistral-large-3:675b': { id: 'mistral-large-3:675b', name: 'Mistral Large 3 675B', maxTokens: 16384, contextWindow: 131072 },
+  'gpt-oss:120b': { id: 'gpt-oss:120b', name: 'GPT-OSS 120B', maxTokens: 16384, contextWindow: 131072 },
   'minimax-m2.5': { id: 'minimax-m2.5', name: 'MiniMax M2.5', maxTokens: 16384, contextWindow: 131072 },
+  'glm-5': { id: 'glm-5', name: 'GLM 5', maxTokens: 16384, contextWindow: 128000 },
   'glm-4.7': { id: 'glm-4.7', name: 'GLM 4.7', maxTokens: 16384, contextWindow: 128000 },
-  // General purpose / large context fallback
-  'kimi-k2.5': { id: 'kimi-k2.5', name: 'Kimi K2.5', maxTokens: 16384, contextWindow: 256000 },
-  'llama4': { id: 'llama4', name: 'Llama 4 Scout', maxTokens: 16384, contextWindow: 131072 },
-  'qwen3:30b-a3b': { id: 'qwen3:30b-a3b', name: 'Qwen3 30B A3B', maxTokens: 16384, contextWindow: 131072 },
+  // Small / fast models
+  'gemma3:27b': { id: 'gemma3:27b', name: 'Gemma 3 27B', maxTokens: 8192, contextWindow: 131072 },
+  'gemma3:12b': { id: 'gemma3:12b', name: 'Gemma 3 12B', maxTokens: 8192, contextWindow: 131072 },
+  'gemma3:4b': { id: 'gemma3:4b', name: 'Gemma 3 4B', maxTokens: 8192, contextWindow: 131072 },
+  'ministral-3:14b': { id: 'ministral-3:14b', name: 'Ministral 3 14B', maxTokens: 8192, contextWindow: 131072 },
+  'ministral-3:8b': { id: 'ministral-3:8b', name: 'Ministral 3 8B', maxTokens: 8192, contextWindow: 131072 },
+  // Special
+  'gemini-3-flash-preview': { id: 'gemini-3-flash-preview', name: 'Gemini 3 Flash Preview', maxTokens: 16384, contextWindow: 131072 },
 };
 
 class OllamaProvider extends BaseProvider {
@@ -29,7 +43,7 @@ class OllamaProvider extends BaseProvider {
       name: 'ollama',
       baseUrl: config.baseUrl || 'https://ollama.com',
       models: config.models || OLLAMA_MODELS,
-      defaultModel: config.defaultModel || 'qwen3-coder',
+      defaultModel: config.defaultModel || 'qwen3-coder:480b',
       ...config,
     });
     this.timeout = config.timeout || 180000;

package/cli/providers/registry.js

@@ -18,9 +18,9 @@ const { checkBudget } = require('../costs');
 // Used during fallback to pick an equivalent model on a different provider.
 
 const MODEL_EQUIVALENTS = {
-  top:    { ollama: 'kimi-k2.5',   openai: 'gpt-4.1',      anthropic: 'claude-sonnet-4-5', gemini: 'gemini-2.5-pro' },
-  strong: { ollama: 'qwen3-coder', openai: 'gpt-4o',        anthropic: 'claude-sonnet',     gemini: 'gemini-2.5-flash' },
-  fast:   { ollama: 'devstral',    openai: 'gpt-4.1-mini',  anthropic: 'claude-haiku',      gemini: 'gemini-2.0-flash' },
+  top:    { ollama: 'kimi-k2.5',        openai: 'gpt-4.1',      anthropic: 'claude-sonnet-4-5', gemini: 'gemini-2.5-pro' },
+  strong: { ollama: 'qwen3-coder:480b', openai: 'gpt-4o',        anthropic: 'claude-sonnet',     gemini: 'gemini-2.5-flash' },
+  fast:   { ollama: 'devstral-small-2:24b', openai: 'gpt-4.1-mini', anthropic: 'claude-haiku',   gemini: 'gemini-2.0-flash' },
 };
 
 // Reverse lookup: model → tier
@@ -136,9 +136,13 @@ function getActiveModel() {
  */
 function parseModelSpec(spec) {
   if (!spec) return { provider: null, model: null };
-  const parts = spec.split(':');
-  if (parts.length >= 2) {
-    return { provider: parts[0], model: parts.slice(1).join(':') };
+  const colonIdx = spec.indexOf(':');
+  if (colonIdx > 0) {
+    const prefix = spec.slice(0, colonIdx);
+    // Only treat as provider:model if prefix is a known provider name
+    if (providers[prefix] || ['ollama', 'openai', 'anthropic', 'gemini', 'local'].includes(prefix)) {
+      return { provider: prefix, model: spec.slice(colonIdx + 1) };
+    }
   }
   return { provider: null, model: spec };
 }

package/cli/sub-agent.js

@@ -48,6 +48,7 @@ function isRetryableError(err) {
   return false;
 }
 
+<<<<<<< Updated upstream
 async function callWithRetry(messages, tools, options) {
   let lastError;
   for (let attempt = 0; attempt <= MAX_CHAT_RETRIES; attempt++) {
@@ -55,6 +56,13 @@ async function callWithRetry(messages, tools, options) {
       // Use callStream (stream:true) — more reliable than callChat (stream:false)
       // with Ollama Cloud. Silently collect the full response via no-op onToken.
       return await callStream(messages, tools, { ...options, onToken: () => {} });
+=======
+async function callChatWithRetry(messages, tools, options) {
+  let lastError;
+  for (let attempt = 0; attempt <= MAX_CHAT_RETRIES; attempt++) {
+    try {
+      return await callChat(messages, tools, options);
+>>>>>>> Stashed changes
     } catch (err) {
       lastError = err;
       if (attempt < MAX_CHAT_RETRIES && isRetryableError(err)) {
@@ -216,7 +224,11 @@ ERROR RECOVERY:
 
   try {
     for (let i = 0; i < maxIter; i++) {
+<<<<<<< Updated upstream
       const result = await callWithRetry(messages, availableTools, chatOptions);
+=======
+      const result = await callChatWithRetry(messages, availableTools, chatOptions);
+>>>>>>> Stashed changes
 
       // Guard against null/undefined responses
       if (!result || typeof result !== 'object') {
@@ -422,4 +434,8 @@ async function executeSpawnAgents(args) {
   }
 }
 
+<<<<<<< Updated upstream
 module.exports = { runSubAgent, executeSpawnAgents, clearAllLocks, classifyTask, pickModelForTier, resolveSubAgentModel, isRetryableError, callWithRetry };
+=======
+module.exports = { runSubAgent, executeSpawnAgents, clearAllLocks, classifyTask, pickModelForTier, resolveSubAgentModel, isRetryableError, callChatWithRetry };
+>>>>>>> Stashed changes

package/cli/tool-tiers.js

@@ -23,12 +23,28 @@ const TIERS = {
  */
 const MODEL_TIERS = {
   // Ollama Cloud — tier by capability
+  'qwen3-coder:480b': 'full',
+  'qwen3-coder-next': 'full',
   'kimi-k2.5': 'full',
-  'qwen3-coder': 'full',
-  'devstral': 'standard',
-  'deepseek-r1': 'standard',
-  'llama4': 'standard',
-  'qwen3:30b-a3b': 'essential',
+  'kimi-k2:1t': 'full',
+  'kimi-k2-thinking': 'full',
+  'deepseek-v3.2': 'full',
+  'deepseek-v3.1:671b': 'full',
+  'devstral-2:123b': 'full',
+  'devstral-small-2:24b': 'standard',
+  'cogito-2.1:671b': 'full',
+  'qwen3-next:80b': 'full',
+  'qwen3.5:397b': 'full',
+  'mistral-large-3:675b': 'full',
+  'gpt-oss:120b': 'full',
+  'minimax-m2.5': 'full',
+  'glm-5': 'full',
+  'glm-4.7': 'standard',
+  'gemma3:27b': 'standard',
+  'gemma3:12b': 'essential',
+  'gemma3:4b': 'essential',
+  'ministral-3:14b': 'standard',
+  'ministral-3:8b': 'essential',
 
   // OpenAI — all full
   'gpt-4o': 'full',

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "nex-code",
-  "version": "0.3.0",
+  "version": "0.3.2",
   "description": "Nex Code — Agentic Coding CLI with Multi-Provider Support",
   "bin": {
     "nex-code": "./bin/nex-code.js"