Skip to main content
OpenRouter provides unified API access to 100+ AI models from multiple providers including OpenAI, Anthropic, Google, Meta, and more.

​
Prerequisites

  • OpenRouter account: openrouter.ai
  • API key (free credits available)
  • Credits for paid models (or use free models)

​
Setup

1

Get API Key

  1. Sign up at openrouter.ai
  2. Navigate to Keys in your account
  3. Click Create Key
  4. Copy the key (starts with sk-or-)
OpenRouter provides $1 in free credits to test models.
2

Open Glyph AI Settings

Go to Settings → AI and select the OpenRouter profile.
3

Add API Key

  1. Click Set API Key in the authentication section
  2. Paste your OpenRouter API key
  3. Click Save
The key is stored in .glyph/app/ai_secrets.json in your space directory.
4

Select Model

Click the Model dropdown. Glyph fetches 100+ available models from OpenRouter.Popular models:
  • anthropic/claude-3.5-sonnet
  • openai/gpt-4o
  • google/gemini-pro-1.5
  • meta-llama/llama-3.1-70b-instruct
  • mistralai/mistral-large
Free models are marked in the OpenRouter model list.
5

Test Connection

Open the AI panel and send a test message. You should receive a response from your selected model.

​
Configuration

​
Provider Settings

  • Service: openrouter
  • Base URL: https://openrouter.ai/api/v1 (default)
  • Authentication: Bearer token (API key)

​
Model Routing

OpenRouter automatically routes your request to the best available provider for your selected model.

​
Model Selection

Glyph fetches the latest model list from OpenRouter’s /v1/models endpoint.

​
Model Details

OpenRouter provides rich model metadata:
  • Context length - Maximum tokens per request
  • Pricing - Input and output token costs
  • Modalities - Text, image, audio support
  • Max output tokens - Maximum response length
  • Supported parameters - Available model options
Hover over the info icon in the model selector to view these details.

​
Free Models

ModelProviderUse Case
meta-llama/llama-3.1-8b-instruct:freeMetaGeneral purpose
mistralai/mistral-7b-instruct:freeMistralFast tasks
google/gemini-flash-1.5:freeGoogleMultimodal
Free models have rate limits but are great for testing and light usage.
ModelInput CostOutput Cost
anthropic/claude-3.5-sonnet$3/1M tokens$15/1M tokens
openai/gpt-4o$2.50/1M tokens$10/1M tokens
google/gemini-pro-1.5$1.25/1M tokens$5/1M tokens
meta-llama/llama-3.1-70b-instruct$0.88/1M tokens$0.88/1M tokens
Check current pricing at openrouter.ai/models.

​
Features

​
Chat Mode

Conversational interaction:
  • Back-and-forth dialogue with any model
  • No file system access
  • Fast responses
  • Best for discussion and brainstorming

​
Create Mode

AI with workspace tools:
  • read_file - Read files from your space
  • search_notes - Search note content
  • list_dir - List directory contents
  • Tool usage tracked in timeline view
  • Best for research and knowledge retrieval
Tool support varies by model. Most modern models (GPT-4, Claude, Gemini) support function calling.

​
Context Attachment

Attach notes for grounded responses:
  • Attach files or folders via context menu
  • Mention with @filename syntax
  • Configure character budget (up to 250K chars)
  • Context limits vary by model

​
API Usage and Billing

​
Credits

OpenRouter uses a credit system:

​
Cost Estimation

Use the model selector to view pricing before sending:
  1. Select a model
  2. Hover over the info icon
  3. View input/output pricing
  4. Estimate cost: (input_tokens × input_price) + (output_tokens × output_price)

​
Rate Limits

Rate limits depend on your credit balance and model:

​
Troubleshooting

​
”API key not set for this profile”

Solution: Add your OpenRouter API key in Settings → AI.

​
”model list failed (401)”

Solution: Your API key is invalid. Generate a new key from OpenRouter.

​
”insufficient credits”

Solution: Add credits at openrouter.ai/credits or use a free model.

​
”model not found”

Cause: The model ID is incorrect or the model is no longer available. Solution: Refresh the model list or check openrouter.ai/models for valid model IDs.

​
Model list is very long

Solution: Use the search/filter in the model selector. Type to filter by provider or model name.

​
Responses from different providers

Cause: OpenRouter routes to the best available provider for your model. If primary provider is unavailable, it uses a fallback. Solution: This is expected behavior. Check OpenRouter’s dashboard for routing details.

​
Advanced Features

​
Provider Preferences

Some OpenRouter models allow provider selection (e.g., anthropic/claude-3.5-sonnet:beta vs anthropic/claude-3.5-sonnet). Use the full model ID from openrouter.ai/models to specify routing.

​
Custom Headers

OpenRouter supports custom headers for advanced use cases: 
[
  { "key": "HTTP-Referer", "value": "https://yourdomain.com" },
  { "key": "X-Title", "value": "Glyph Integration" }
]
These headers may provide better rate limits and routing.

​
Security Best Practices

  • Never commit .glyph/app/ai_secrets.json to version control
  • Rotate API keys if exposed
  • Monitor usage at openrouter.ai/activity
  • Set spending limits in your OpenRouter account

​
OpenRouter vs Direct Providers

​
Advantages of OpenRouter

✅ Single API key for 100+ models ✅ Automatic provider failover ✅ Unified billing ✅ Model comparison and discovery ✅ Free models available

​
When to Use Direct Providers

  • Lower latency (no routing layer)
  • Provider-specific features (e.g., Anthropic’s thinking models)
  • Direct billing relationship
  • Higher rate limits with provider-native keys

​
Next Steps

Chat Modes

Learn about chat vs create modes

Context Management

Attach notes to conversations

OpenAI

Compare with direct OpenAI integration

Profiles

Manage multiple AI profiles

Build docs developers (and LLMs) love

Get started for freeTalk to us