Skip to content

Add Your API Key

RagLeap uses your own AI API key — this keeps costs transparent and your data private.

Supported Providers

Provider Models Get Key
Google Gemini gemini-2.5-flash, gemini-1.5-flash aistudio.google.com
OpenAI gpt-4.1, gpt-4o-mini platform.openai.com
Anthropic claude-opus-4-6, claude-haiku console.anthropic.com
Mistral mistral-large, mistral-small console.mistral.ai
OpenRouter 100+ models openrouter.ai
Custom / Ollama Any OpenAI-compatible endpoint + local models Your endpoint or localhost

Setup Steps

  1. Go to Settings → AI Settings
  2. Enable Use Your Own API Key
  3. Select your provider
  4. Paste your API key
  5. Select your model
  6. Save

Embedding Key (Required)

A separate embedding key is needed for document uploads.

Recommended: Gemini embedding key (free)

  1. Go to aistudio.google.com
  2. Create a free API key
  3. Paste in Settings → AI Settings → Embedding Key
Use Case Model Why
Daily use gemini-2.5-flash Free, fast, 1M TPM
Deep analysis claude-opus-4-6 Best reasoning
Balanced gpt-4o-mini Good performance, low cost

Using Local AI Models (Ollama)

RagLeap works with any OpenAI-compatible API — including Ollama running on your own machine or server. This gives you 100% offline AI with zero API costs.

Step 1 — Install Ollama

Linux / Mac: 

curl -fsSL https://ollama.com/install.sh | sh

Windows: Download from ollama.com

Step 2 — Pull a model

ollama pull llama3.2        # 2GB — fast, good quality
ollama pull mistral         # 4GB — excellent reasoning
ollama pull qwen2.5:7b      # 4GB — great multilingual
ollama pull phi4-mini       # 2GB — lightweight, fast

Step 3 — Start Ollama server

ollama serve
# Now running at http://localhost:11434

For remote access (self-hosted RagLeap on a different server): 

OLLAMA_HOST=0.0.0.0:11434 ollama serve

Step 4 — Connect to RagLeap

  1. Go to Settings → AI Settings
  2. Select provider: Custom
  3. Fill in:
Field Value
API URL http://localhost:11434/v1
API Key ollama (any string works)
Model llama3.2 (or any pulled model)
  1. Save — done!
Use Case Model Size Command
Fast responses llama3.2 2GB ollama pull llama3.2
Best quality mistral 4GB ollama pull mistral
Multilingual qwen2.5:7b 4GB ollama pull qwen2.5:7b
Low RAM (2GB) phi4-mini 2GB ollama pull phi4-mini
Code tasks codellama 4GB ollama pull codellama

Self-Hosted RagLeap + Ollama = Zero API costs

Running RagLeap Server Edition with Ollama means zero ongoing costs. Your data never leaves your server. Perfect for hospitals, banks, and law firms.

Embedding with Ollama

For document embeddings with Ollama: - API URL: http://localhost:11434/v1 - Model: nomic-embed-text - Run: ollama pull nomic-embed-text

Keep your key private

Never share your API key. RagLeap encrypts it at rest and never exposes it to customers.