JoinGrove

Ollama Integration

Source: docs/integrations/ollama.md
Generated: 01/09/2026

Connect JoinGrove to Ollama for local coding models — no cloud subscription required.

Key Concepts

  • What: Ollama serves local LLMs; JoinGrove preconfigures endpoints, BYOK routing, and adaptive model tiers in product.json.
  • Why: Global South developers need offline-capable AI on 8 GB laptops and campus lab servers.
  • Related: Full tutorial (offline-ai-ollama.md), models overview (overview.md), offline providers reference (offline-providers.md).

API / Reference

Resource Value
Default URL http://localhost:11434
Lite model (8 GB RAM) qwen2.5-coder:1.5b
Standard model (16 GB+) qwen2.5-coder:7b
Config file ~/.joingrove/User/chatLanguageModels.json
Product config product.json → joingroveOfflineProviders

Commands

Command ID
Set up local AI joingrove.setupOfflineAi
Check connection joingrove.checkOllamaStatus
Open offline AI docs joingrove.openOfflineAiDocs
Setup script npm run setup-offline-ai

Setup & Installation

  1. Install Ollama from ollama.com and run ollama serve.
  2. Run npm run setup-offline-ai or Command Palette → Set Up Local AI (Ollama).
  3. Verify with Check Ollama Connection (joingrove.checkOllamaStatus).
  4. Select your Ollama model in the Agents chat model picker.

Code Examples

Pull and connect a coding model:

ollama pull qwen2.5-coder:7b
npm run setup-offline-ai

Remote lab server:

npm run setup-offline-ai -- --url http://192.168.1.10:11434

Dry run:

npm run setup-offline-ai -- --dry-run

Next: Open offline-providers.md for the full model and RAM tier table.

How It Connects