AI Provider Configuration

Redline supports multiple AI providers for the research assistant. Configure your preferred provider in Settings > AI Provider.

Supported Providers

Provider	Vision	Tools	Local	Notes
Redline AI	Yes	Yes	No	Included with license, no API key needed
Claude (Anthropic)	Yes	Yes	No	Best overall with own API key
OpenAI	Yes	Yes	No	Strong alternative
OpenRouter	Varies	Varies	No	Access many models
Ollama	Varies	Varies	Yes	Fully local, no API key

Redline AI

Redline AI is the managed AI service included with your Redline license. It provides access to powerful AI models without needing to manage your own API keys.

Features

• No API key required - Works automatically with your license
• Included usage budget - Monthly AI credits based on your tier
• Curated models - Access to top-performing models for research
• Zero configuration - Just select Redline AI and start working

Setup

1. Go to Settings > AI Provider
2. Select Redline AI
3. That's it! Your license key handles authentication automatically.

Available Models

Model	Best For
Claude Sonnet 4	General research, best for tool use
Claude 3.5 Haiku	Fast responses, cost-effective
Kimi K2.5	Alternative for variety

Usage Budget

Your monthly AI budget depends on your license tier:

Tier	Monthly Budget
Personal	$2
Pro	$8
Enterprise	$8 per seat

Budget resets monthly from your license activation date. When budget is exhausted, AI features pause until the next month.

Checking Usage

• In-app: Settings > AI Provider shows remaining budget
• Enterprise: Use the Usage Dashboard for detailed tracking

Budget Tips

• Use Claude 3.5 Haiku for routine tasks (cheaper per token)
• Use Claude Sonnet 4 for complex research
• Limit context by selecting fewer nodes when chatting

When to Use Your Own API Key

You might prefer your own API key if you:

• Need unlimited usage beyond your budget
• Want access to specific models not in Redline AI
• Have organizational API key requirements
• Prefer to run models locally (Ollama)

You can switch between Redline AI and other providers anytime in Settings.

---

Claude (Anthropic)

Claude is the recommended provider for Redline. It offers excellent research capabilities and tool use.

Authentication Options

1. OAuth Login (Easiest)

1. Go to Settings > AI Provider
2. Select "Claude"
3. Click "Sign in with Claude"
4. Authorize Redline in your browser
5. You're connected!

2. API Key

1. Get an API key from console.anthropic.com
2. Go to Settings > AI Provider
3. Select "Claude"
4. Enter your API key
5. Click "Save"

Recommended Models

Model	Best For
claude-sonnet-4-20250514	General use, best balance
claude-3-5-haiku-20241022	Fast, cost-effective

Token Limits

Model	Context Window
Claude Sonnet 4	200K tokens
Claude 3.5 Haiku	200K tokens

---

OpenAI

OpenAI provides GPT-4 and other models with strong capabilities.

Setup

1. Get an API key from platform.openai.com
2. Go to Settings > AI Provider
3. Select "OpenAI"
4. Enter your API key
5. Click "Save"

Recommended Models

Model	Best For
gpt-4-turbo	Best capability
gpt-4o	Fast, vision support
gpt-3.5-turbo	Cost-effective

---

OpenRouter

OpenRouter provides access to many models through a single API, including models from Anthropic, OpenAI, Meta, Google, and more.

Setup

1. Get an API key from openrouter.ai
2. Go to Settings > AI Provider
3. Select "OpenRouter"
4. Enter your API key
5. Select a model from the list
6. Click "Save"

Available Models

OpenRouter provides access to 100+ models. Some popular options:

Model	Provider	Notes
claude-3-5-sonnet	Anthropic	Via OpenRouter
gpt-4-turbo	OpenAI	Via OpenRouter
llama-3.1-70b	Meta	Open weights
gemini-pro	Google	Google's model

Pricing

OpenRouter charges per token based on the model used. Check openrouter.ai/models for current pricing.

---

Ollama (Local)

Ollama runs AI models locally on your machine. No API key or internet required for AI inference.

Setup

1. Install Ollama from ollama.ai
2. Pull a model:

   ollama pull llama3.1

3. Start Ollama (it runs as a background service)
4. Go to Settings > AI Provider
5. Select "Ollama"
6. Select your model
7. Click "Save"

Recommended Models

Model	Size	Notes
llama3.1:8b	~5GB	Good for most machines
llama3.1:70b	~40GB	Best quality, needs GPU
mistral:7b	~4GB	Fast, efficient
mixtral:8x7b	~26GB	Strong performance

Hardware Requirements

Model Size	RAM	GPU VRAM
7B	8GB+	6GB+
13B	16GB+	10GB+
70B	64GB+	40GB+

Performance

For best Ollama performance, use a GPU with sufficient VRAM. CPU inference is slower but works.

Custom Ollama URL

If Ollama runs on a different machine or port:

1. Go to Settings > AI Provider > Ollama
2. Enter the custom URL (e.g., http://192.168.1.100:11434)
3. Save and verify connection

---

Provider Comparison

Capability Matrix

Feature	Redline AI	Claude	OpenAI	OpenRouter	Ollama
Tool Use	Yes	Yes	Yes	Model-dependent	Model-dependent
Vision	Yes	Yes	Yes	Model-dependent	Some models
Long Context	Yes	Yes	Yes	Model-dependent	Limited
Streaming	Yes	Yes	Yes	Yes	Yes
Offline	No	No	No	No	Yes
API Key Required	No	Yes	Yes	Yes	No

Cost Comparison

Provider	Approximate Cost
Redline AI	Included with license ($2-8/month)
Claude Sonnet	$3/$15 per 1M tokens (in/out)
GPT-4 Turbo	$10/$30 per 1M tokens
OpenRouter	Varies by model
Ollama	Free (local compute)

Recommendations

• Easiest setup: Redline AI (included, no configuration)
• Unlimited usage: Claude or OpenAI (with your API key)
• Privacy-focused: Ollama (fully local)
• Model variety: OpenRouter
• OpenAI ecosystem: OpenAI

---

Redline AI for Enterprise

Enterprise licenses get enhanced Redline AI features:

• Per-seat budgets - Each seat gets $8/month
• Usage dashboard - Track usage across all seats at getredline.io/usage
• Centralized management - Monitor team usage from one place

Enterprise users can still configure their own API keys if they prefer to bypass Redline AI.

See Enterprise Usage Dashboard for more details.

---

Troubleshooting

"Invalid API Key"

• Verify the key is correct (no extra spaces)
• Check the key hasn't expired
• Ensure you have API access enabled on your account

"Model not available"

• The model may be deprecated
• Check you have access to that model tier
• Try a different model

"Rate limit exceeded"

• Wait a few minutes and retry
• Consider upgrading your API tier
• Reduce request frequency

Ollama not connecting

• Verify Ollama is running: ollama list
• Check the URL is correct (default: http://localhost:11434)
• Ensure no firewall is blocking the connection

Slow responses

• Try a smaller/faster model
• Check your internet connection
• For Ollama, verify GPU is being used