Fusion Studio

A lightweight chat interface for OpenRouter AI models. Manage up to 10 concurrent chat threads, load custom prompts, and integrate JSON data seamlessly.

Getting Started

How to Get an API Key from OpenRouter

Visit OpenRouter.ai
Sign up or log in to your account
Navigate to your Dashboard or API Keys section
Click Create API Key
Give your key a name (e.g., "Fusion Studio")
Copy your new API key (it will look something like sk-or-v1-...)
Keep this key safe! You'll need it to use Fusion Studio

Note: OpenRouter requires credits to use. Make sure your account has a balance before making API calls.

Setting Your API Key

Open Fusion Studio in your browser
Click the Add API Key button (located in the top header, to the right of the model selector)
Paste your OpenRouter API key into the prompt
Click OK or press Enter

Your API key is saved locally in your browser and will persist between sessions. The button will change to Clear API Key once a key is set.

IMPORTANT: If your laptop is ever lost or stolen, IMMEDIATELY go to openrouter.ai and revoke your API key. You should also consider setting a spend limit as an extra layer of protection.

Using the Assistant Chat

Fusion Studio includes a built-in Assistant that runs locally in your browser using the Phi-3-mini model. This means you can use it without an API key!

To access the Assistant:

Look at the left panel at the very top
Click the 💬 Assistant button
Start asking questions about how to use Fusion Studio

The Assistant is perfect for:

Learning how Fusion Studio works
Getting help with features
Understanding API keys and chat interfaces
No API key required!

Starting a Thread

Fusion Studio gives you 10 chat slots so you can run multiple conversations at once.

To start a new thread:

Look at the left panel with numbered slots (1–10)
Click any slot that says "Empty..."
A prompt will appear asking you to name your chat
Type a name and click OK
The slot will now show your chat name

To switch between threads:

Click any chat slot to switch to that conversation
The active chat name appears at the top of the screen

To manage a thread:

Right-click on any chat slot to open the context menu
Choose from:
- Rename – Change the chat name
- Export Chat – Download the conversation as JSON
- Clear Data – Delete all messages (keeps prompt, model, and JSON)

Setting and Saving a Prompt

The right panel is your prompt area. This is where you load a system prompt that tells the AI how to behave.

To create or edit a prompt:

The prompt editor is visible in the right panel
Click the Edit button
Type your prompt directly in the text area, OR click Add File to load a .md (Markdown) file
Click Save when finished

To edit an existing prompt:

Click the Edit button
Make your changes in the text area
Click Save to apply changes

To load a prompt from a file:

Click Edit, then Add File
Select a .md (Markdown) file from your computer
The file contents will appear in the prompt text area
Click Save to apply

PLEASE NOTE: Prompt applies only to the current chat. You will need to re-upload or copy paste your prompt when starting another chat.

Why use prompts?
Prompts give the AI personality, instructions, or context. For example:

"You are a helpful coding assistant."
"Answer all questions in pirate speak."
"You are an expert in JavaScript debugging."

Adding a JSON File

Fusion Studio lets you attach JSON files to your messages. This is useful for feeding data into the AI or working with structured information.

To add a JSON file:

Click the Add File button (located in the input footer, bottom-left)
Select a .json file from your computer
The filename will appear where it said "No file loaded"
The button will change to Remove File
Type your message in the text area
Click Send

What happens:

The JSON file contents are included with your message
The AI can read and work with the data
The file stays loaded until you remove it or switch chats

To remove a JSON file:

Click the Remove File button and confirm removal

Managing Your Messages

You can edit, regenerate, or delete any of your previous messages using the right-click context menu.

To access message options:

Right-click on any of your messages (user bubbles in the chat)
Choose from:
- Regenerate – Deletes this message and everything after it, then resends the message
- Delete – Deletes this message and everything after it permanently
- Edit – Deletes this message and everything after, and places the text in the input field for editing

Warning: All three options will delete the selected message and ALL messages that came after it. This cannot be undone!

Why use this?

Fix typos in past messages
Try different phrasings to get better responses
Remove unwanted conversation branches
Start fresh from any point in the conversation

Exporting Thread Data

You can export any chat thread as a JSON file to save your conversation history.

To export a thread:

Right-click on the chat slot you want to export
Select Export Chat from the context menu
A JSON file will download to your computer

What's in the export?
The JSON file contains:

Chat name
Selected model
Full conversation history (all messages and responses)
Timestamp of export

Why export?

Backup important conversations
Share conversations with others
Analyze chat data programmatically
Import into other tools

Additional Features

Model Selection

Choose from multiple AI models using the dropdown at the top (next to the chat name):

🏠 Phi-3-mini (Local) – Runs in your browser, no API key needed!
DeepSeek models (with extended reasoning)
Qwen models
Inclusion AI models (Ling and Ring)
Moonshot AI models
Plus many more models from OpenRouter!

How to use the model selector:

Click the dropdown to see all available models
OR start typing to search/filter models in real-time
Pro tip: Type "free" (space) "other term" to filter only free models on OpenRouter
Use arrow keys to navigate, Enter to select

Reasoning Display:
Models that support reasoning (like DeepSeek) will show their thinking process in a separate section above the main response. The thinking section is collapsible - it expands automatically as reasoning streams, then auto-collapses when the main response begins.

Markdown Formatting:
All AI responses support markdown formatting including bold, italic, code blocks, lists, and more. The formatting is rendered automatically in both responses and reasoning traces.

Clear All Data

The Clear All button at the bottom of the left panel deletes all 10 chat threads at once, including all messages, prompts, and settings. Use with caution!

Tips & Tricks

Start with the Assistant: New to Fusion Studio? Click the 💬 Assistant button at the top of the left panel
Organize your chats: Use descriptive names like "Debug Session" or "Brainstorm Ideas"
Save your prompts: Keep useful prompts in .md files for quick loading
Try the local model: No API key? Use the 🏠 Phi-3-mini (Local) model
Export frequently: Back up important conversations before they get too long

Troubleshooting

API calls aren't working:

Make sure your API key is set correctly (click Add API Key in the top header)
Check that your OpenRouter account has credits
Verify you have internet connection
Try using the 🏠 Phi-3-mini (Local) model instead (no API key needed)

Chat slots aren't saving:

Make sure you're not in incognito/private browsing mode
Check if your browser allows local storage

JSON file won't load:

Ensure the file is valid JSON (use a validator like jsonlint.com)
Make sure the file extension is .json

About Fusion Studio

Fusion Studio is a browser-based chat client built for power users who need:

Multiple parallel conversations
Custom system prompts
JSON data integration
Full conversation exports
Local AI models (no API key required for Phi-3-mini)

Built with vanilla HTML, CSS, and JavaScript. No frameworks, no bloat.

Enjoy chatting! 🚀