How It Works (Simple)
Let's walk through exactly how Unbrowse works, from installation to execution — in simple terms anyone can understand.
Visual Overview
The Four-Step Process
Step 1: You Browse, We Capture
What you do: Install the Unbrowse browser extension and browse the web normally.
What happens behind the scenes:
While you visit websites and interact with them (logging in, searching, purchasing, posting), the extension quietly observes the network requests your browser sends and receives.
It doesn't record:
Screenshots
Video of your screen
Your passwords or sensitive data
It does capture:
The pattern of HTTP requests (the "language" websites speak)
The sequence of interactions
The structure of requests and responses
Think of it like this: The extension learns "how to speak" to each website you visit.
Step 2: We Turn It Into "Abilities"
What we do: Our AI processes these captured patterns.
What gets created: Reusable "abilities" — step-by-step recipes for how to perform specific actions on specific websites.
Examples of abilities:
"Search for flights on United Airlines"
"Post an update to LinkedIn with an image"
"Check account balance on Chase Bank"
"Add item to Amazon cart"
"Create a Jira ticket with specific priority"
Each ability is like a railroad track: once one person lays it down, everyone can use it forever.
Step 3: Agents Search and Execute
What agents do: When an AI agent needs to do something, it:
Searches the ability index ($0.0001 per search)
"Find ability to book flights"
Returns matching abilities from our index
Executes the chosen ability ($0.001 per execution)
Sends the network request directly to the website
Bypasses all the visual UI
Gets structured results back
From the agent's perspective:
No screenshots. No clicking. No waiting for page loads. Just pure execution.
Step 4: Everyone Gets Paid
What happens with the fees:
When an agent executes an ability, the fee is automatically split:
For searches ($0.0001):
97% → FDRY Treasury (buyback + burn)
3% → Infrastructure costs
For executions ($0.001):
50% → Website owner (they provide the service)
46% → FDRY Treasury (buyback + burn)
3% → You (the indexer who created the ability)
1% → Infrastructure costs
This happens automatically, instantly, and transparently through x402 smart contracts.
A Complete Example: Flight Booking
Let's walk through a real scenario end-to-end.
Preparation Phase
You (the indexer):
Install Unbrowse extension
Visit United Airlines website
Search for a flight normally (Tokyo to New York)
Complete the search flow
Unbrowse:
Captures the network pattern
AI analyzes the requests
Creates "Search United Airlines flights" ability
Stores it in the index
This happens once. You've now created a reusable ability.
Execution Phase (Days/Weeks Later)
An AI agent needs: "Find me flights from Tokyo to New York, December 15-20"
Step 1: Search ($0.0001)
Agent queries Unbrowse index: "Search flights Tokyo to New York"
Unbrowse returns: "United Airlines flight search" ability (among others)
Step 2: Execute ($0.001)
Agent chooses United Airlines ability
Unbrowse sends network request directly to United's API
United responds with flight data
Agent receives structured results in 0.3 seconds
Payment Distribution:
United Airlines: $0.0005 (50%)
FDRY Buyback: $0.00046 (46%, split 50/50 burn/treasury)
You (indexer): $0.00003 (3%)
Infrastructure: $0.00001 (1%)
You earn $0.00003 every time any agent uses the ability you created.
At Scale
If your ability is used 10,000 times per month:
You earn: $0.00003 × 10,000 = $0.30/month
If you created 100 popular abilities averaging 10,000 uses each:
You earn: $30/month in passive income
If some abilities become extremely popular (100,000+ uses):
You could earn hundreds or thousands per month
What Makes This Special
For Users (Indexers)
You don't need:
Technical skills
To write code
To understand APIs
Special equipment
You just:
Browse normally
Earn passive income
First-movers earn most
For Agents
Traditional approach:
Agent wants to book flight:
→ Open browser (5 seconds)
→ Load website (3 seconds)
→ Click search (2 seconds)
→ Wait for results (8 seconds)
→ Parse visual layout (10 seconds)
→ Extract data (5 seconds)
→ Total: ~33 seconds
Success rate: 75%Unbrowse approach:
Agent wants to book flight:
→ Query index (0.01 seconds)
→ Execute ability (0.3 seconds)
→ Receive structured data (instant)
→ Total: ~0.3 seconds
Success rate: 95%+110x faster. 25% more reliable. 50-100x cheaper.
For Websites
Traditional approach:
Agents scrape data without permission
No revenue for website owners
Incentive to block agent traffic
Unbrowse approach:
Websites earn 50% of every agent interaction
Automatic payment via domain verification
Incentive to enable agent access
How Credentials Work
Your Security
Your passwords never leave your browser.
When you browse with your credentials:
Extension captures the pattern of authentication
But not your password itself
Passwords stored locally (encrypted, like 1Password)
When an agent executes:
Uses the agent's own credentials (not yours)
Or your credentials if you explicitly authorize
All encryption happens client-side
We cannot see your passwords. Ever.
The Network Effect
As more people use Unbrowse:
More indexers → More abilities created
More abilities → Better coverage of the web
Better coverage → More agents can succeed
More agents → More usage
More usage → More revenue for everyone
More revenue → Higher FDRY value (burns + scarcity)
Higher value → Attracts more indexers
Loop repeats and accelerates
Getting Started
As an Indexer
Browse normally
Earn passive income
As a Developer
Query ability index
Execute abilities
Build reliable agents
As a Token Holder
Buy $FDRY
Benefit from usage-based burns
Watch scarcity increase with usage
Next: Explore Key Concepts to understand the terminology and technical foundations.
Last updated