Create interface for your Open Router Chat/Voice AI Agent and deploy to your website
How to Use Any AI Model on ConvoCore with OpenRouter
ConvoCore isn't locked to a single AI provider. With our custom LLM support, you can connect any model from OpenRouter's catalog of 400+ AI models — including GPT-4, Claude, Llama, Gemini, Mistral, and more.
In this guide, we'll show you exactly how to set this up in under 5 minutes.
What is OpenRouter?
OpenRouter is a unified API gateway that gives you access to hundreds of AI models through a single endpoint. Instead of juggling separate API keys for OpenAI, Anthropic, Google, and others, you use one key to access them all.
Why use OpenRouter with ConvoCore?
- Access 400+ models from one dashboard
- Switch between models without changing your setup
- Automatic fallbacks if a provider goes down
- Transparent, pay-as-you-go pricing
- Use open-source models like Llama or Mistral at lower costs
Step 1: Create Your OpenRouter Account
Head over to openrouter.ai and sign up for a free account.
Step 2: Generate Your API Key
Once logged in:
- Go to Settings → API Keys or visit openrouter.ai/settings/keys directly
- Click "Create Key"
- Give your key a name (e.g., "ConvoCore Agent")
- Optionally set a credit limit to control spending
- Click create and copy your API key immediately — you won't be able to see it again!
Important: Treat your API key like a password. Never share it publicly or commit it to code repositories.
Step 3: Add Credits to Your OpenRouter Account
Before your API key will work, you need to add credits:
- Go to Account → Settings
- Scroll to the Credits section
- Click "Add Credits" and add your preferred amount
Even $5-10 is enough to get started and test your agent.
Step 4: Choose Your Model
OpenRouter uses a simple naming format for models:
provider/model-name
Popular examples:
| Model | OpenRouter ID |
|---|---|
| GPT-4o | openai/gpt-4o |
| GPT-4 Turbo | openai/gpt-4-turbo |
| Claude 4 Sonnet | anthropic/claude-sonnet-4 |
| Claude 3.5 Sonnet | anthropic/claude-3.5-sonnet |
| Llama 3.1 70B | meta-llama/llama-3.1-70b-instruct |
| Gemini 2.5 Flash | google/gemini-2.5-flash-preview |
| Mistral Large | mistralai/mistral-large |
| DeepSeek V3 | deepseek/deepseek-chat |
Browse the full list at openrouter.ai/models to find the perfect model for your use case.
Step 5: Set Up Your ConvoCore Agent
But first..
What is ConvoCore.ai?
ConvoCore is an all-in-one AI voice and chat agent platform that lets you build, deploy, and manage intelligent conversational agents — without writing code.
What makes ConvoCore different:
- Voice + Chat in one platform — Deploy agents that can handle both text conversations and real phone calls
- Custom LLM support — Use any AI model you want (which is exactly what we're doing in this guide)
- White-label ready — Fully brandable widgets that match your website
- Built-in analytics — Track every conversation, measure performance, and gain insights
- One-click deployment — Go live on any website with a single script tag
- Real phone numbers — Connect your agent to an actual phone line for voice calls
Whether you're building a customer support bot, lead qualification agent, appointment scheduler, or AI receptionist — ConvoCore gives you the infrastructure to deploy fast and scale effortlessly.
Let's connect OpenRouter to your ConvoCore agent. Follow along step by step.
5.1 — Create Your Account
Head to convocore.ai/signup and create your free account.
Complete the quick onboarding flow to set up your workspace.
5.2 — Create Your Agent
Once onboarding is complete, you'll land on your dashboard. Click "Create Agent" to build your first AI agent.
Give your agent a name (e.g., "Customer Support Bot") and continue.
5.3 — Open the Prompt Page
Inside your new agent, navigate to the Prompt tab. This is where all the magic happens — you'll configure your LLM and customize your agent's behavior here.
5.4 — Configure Your Custom LLM
Scroll to the LLM section within the Prompt page:
- Click the LLM Provider dropdown
- Select "Custom"
- Fill in the three fields with your OpenRouter details:
| Field | What to Enter |
|---|---|
| API Key | Your OpenRouter API key (starts with sk-or-...) |
| LLM Endpoint | https://openrouter.ai/api/v1 |
| Model | Your chosen model ID (e.g., anthropic/claude-sonnet-4) |
5.5 — Customize Your Agent's Instructions
Still in the Prompt tab, you'll find the Instructions section (also called the system prompt). This is where you tell your agent who it is and how to behave.
For our example, we want a customer support agent for ConvoCore that always speaks English:You are a friendly and helpful customer support agent for ConvoCore.ai — an AI voice and chat agent platform.Your job is to:
Answer questions about ConvoCore's features and pricing Help users troubleshoot issues with their agents Guide users through setup and configuration Always respond in English, regardless of what language the user writes in Be concise, helpful, and professional. If you don't know something, say so and offer to connect them with our support team.
Paste this into your Instructions field (or write your own based on your use case).
5.6 — Test Your Agent
Before going live, let's make sure everything works.
On the right side of the screen, click "Show Test" to open the test chat panel. Send a few messages to your agent and verify:
- ✅ It responds correctly
- ✅ It follows your instructions
- ✅ The tone and style match what you want
If something's off, tweak your instructions or check your LLM settings.
Step 6: Deploy Your Agent
Everything working? Let's go live.
In the top right corner, click the little button beside the Publish button then from the dropdown select Website Widget. You'll see your embed code — a simple JavaScript snippet:
Copy that code and paste it into your website's HTML, just before the closing </body> tag.
Works everywhere: WordPress, Webflow, Shopify, Wix, Squarespace, Framer, custom HTML — if it supports custom code, ConvoCore works. The chat widget will automatically appear in the bottom right corner of your site. Done! 🎉
Bonus: What You Get with ConvoCore 📊 Full Analytics Dashboard Every ConvoCore agent comes with a built-in analytics dashboard — no extra setup required:
All chat transcripts — Read every conversation your agent has had Usage metrics — Track message volume, response times, and engagement User insights — See who's chatting, when, and from where Export data — Download everything in CSV or JSON for your own analysis
The moment your agent goes live, data starts flowing into your dashboard automatically.
📞 Voice Integration + Real Phone Numbers
Here's where ConvoCore really stands out — you can connect your agent to voice.
Voice widget — Let users talk to your agent directly from your website Real phone numbers — Get an actual phone number that routes to your AI agent Inbound + outbound calls — Handle incoming calls or have your agent make calls Same brain, different channel — Your agent uses the same instructions and LLM for both chat and voice
Turn your chat agent into a full AI receptionist that answers phone calls 24/7.
Troubleshooting Tips
Agent not responding?
Double-check your API key has credits loaded in OpenRouter Verify the model ID is exactly correct (case-sensitive) Make sure the endpoint is https://openrouter.ai/api/v1 — no trailing slash, make sure there are no typos in the model name at all.
Responses are slow?
Try faster models like google/gemini-2.5-flash-preview or anthropic/claude-3-haiku Check OpenRouter's status page for issues
Widget not showing on your site?
Place the script before , not in Check browser console for errors Temporarily disable ad blockers to test
Want to switch models later?
Just update the Model field — everything else stays the same
Get Started Now
Create your OpenRouter account — Get your API key Sign up for ConvoCore — Build your agent https://convocore.ai/signup Configure — Plug in OpenRouter, write your instructions Test — Make sure it works Deploy — One line of code, you're live Monitor — Watch it all from your dashboard
Total time: Under 10 minutes.
Need Help?
Email: support@convocore.ai Docs: docs.convocore.ai