Write Features In Plain English
Describe what you want in plain English. The compiler checks it against the ontology, generates TypeScript and tests, and deploys to the edge. No coding required.
Write What You Want
The System Builds It
Example: Chat with AI Clone
FEATURE: Let fans chat with my AI clone
WHEN a fan sends a message
CHECK they own tokens
GET my AI personality
SEND message to AI with my personality
SAVE the conversation
GIVE fan 10 tokens as reward
SHOW AI response to fan
That's it. The system builds the entire feature from this—frontend, backend, database, tests, deployment.
Command Reference
Every command maps to a specific operation in the ontology. Simple, powerful, and type-safe.
CREATE ai clone called "John's AI Clone"
WITH voice ID from previous step
WITH personality prompt
WITH active status
CREATE token called "CREATOR"
WITH total supply of 10 million
WITH symbol "CRTR"
CONNECT creator to ai clone as owner
CONNECT fan to token as holder
CONNECT student to course as enrolled
RECORD clone created
BY creator
WITH voice ID and personality
RECORD tokens purchased
BY fan
WITH amount and price
CALL ElevenLabs to clone voice
WITH audio samples from videos
SAVE AS voice ID
CALL Stripe to charge payment
WITH amount $100
SAVE AS payment
CHECK creator exists
OTHERWISE say "Creator not found"
CHECK fan has tokens
OTHERWISE say "Buy tokens to chat"
GET fan's token balance WHERE token is CREATOR
SAVE AS balance
GET all content WHERE creator owns it
SAVE AS my content
WAIT 24 hours
WAIT until payment received
FOR EACH video IN videos
DO extract audio
GIVE clone ID and voice ID
GIVE success message
STREAM "Weekly Q&A" TO youtube
MIX WITH AI CLONE
REFER fan to new user
REWARD 100 tokens
NOTIFY students ABOUT lesson
VIA email and push
Real-World Features
Production-ready examples showing the full power of the Plain English DSL
Create AI Voice Clone
FEATURE: Create my AI voice clone from videos
INPUT:
- creator: which creator this is for
- video links: list of my videos (at least 3)
OUTPUT:
- clone ID: my new AI clone
- voice ID: my cloned voice
FLOW:
CHECK creator exists
OTHERWISE say "Creator not found"
CHECK video links are at least 3
OTHERWISE say "Need at least 3 videos to clone your voice"
CALL OpenAI to extract audio from videos
WITH video links
SAVE AS audio samples
CALL ElevenLabs to clone voice
WITH audio samples
WITH name "Creator's Voice"
RETRY 3 times if it fails
TIMEOUT after 5 minutes
SAVE AS voice ID
CALL OpenAI to analyze personality
WITH video transcripts
WITH social media posts
SAVE AS personality
CREATE ai clone called "Creator's AI Clone"
WITH voice ID
WITH personality prompt
WITH active status
SAVE AS clone ID
CONNECT creator to clone ID as owner
RECORD clone created
BY creator
WITH voice ID
WITH personality traits
GIVE clone ID and voice ID
Services Used
Features
Token Purchase with Atomic Transactions
FEATURE: Let fans buy creator tokens
INPUT:
- fan: who is buying
- token: which creator's token
- amount: how many tokens
- usd amount: price in dollars
OUTPUT:
- payment ID: transaction receipt
- tx hash: blockchain confirmation
FLOW:
CHECK fan exists
OTHERWISE say "Fan account not found"
CHECK token exists
OTHERWISE say "Token not found"
CHECK amount is greater than 0
OTHERWISE say "Amount must be positive"
DO TOGETHER:
- CALL Stripe to charge payment
WITH usd amount in cents
WITH fan's payment method
SAVE AS payment
- CALL Blockchain to mint tokens
WITH token contract address
WITH fan's wallet
WITH amount
SAVE AS mint transaction
IF ANY FAIL:
- CALL Stripe to refund payment
- CALL Blockchain to burn tokens
RECORD tokens purchased
BY fan
WITH amount
WITH usd amount
WITH payment ID
WITH tx hash
GET fan's token balance WHERE token matches
SAVE AS existing balance
IF existing balance exists
THEN UPDATE balance to add amount
ELSE CREATE new balance
CONNECT fan to token as holder
WITH balance amount
GIVE payment ID and tx hash
Ontology Mapping
- Things: fan, token
- Connections: holds_tokens (with balance)
- Events: tokens_purchased
- Services: Stripe, Blockchain
Advanced Features
Chat with AI Clone
FEATURE: Let fans chat with my AI clone
INPUT:
- fan: who is chatting
- clone: which AI clone
- message: what fan said
OUTPUT:
- response: what AI said back
- tokens earned: reward for engagement
FLOW:
CHECK fan exists
OTHERWISE say "Please sign up first"
CHECK clone exists
OTHERWISE say "AI clone not found"
GET fan's token balance WHERE clone's token
SAVE AS balance
CHECK balance is greater than 0
OTHERWISE say "Buy tokens to chat with this creator"
GET last 10 messages
WHERE fan chatted with clone
SAVE AS conversation history
CALL OpenAI to search relevant knowledge
WITH fan's message
WITH clone's knowledge base
SAVE AS context
CALL OpenAI to generate response
WITH clone's personality
WITH conversation history
WITH context
WITH fan's message
SAVE AS AI response
RECORD clone interaction
BY fan
WITH message
WITH response
WITH tokens used
UPDATE fan's token balance
ADD 10 tokens as reward
GIVE AI response and 10 tokens earned
AI Features
Gamification
Production-Ready Patterns
Handle errors, retry failures, and ensure data consistency with battle-tested patterns
TRY:
CALL payment processor
CALL blockchain
IF IT FAILS:
CALL refund payment
SEND error notification
ALWAYS DO:
RECORD what happened
SAVE logs
DO AT SAME TIME:
- CALL OpenAI to analyze videos
- CALL OpenAI to analyze posts
- CALL OpenAI to analyze comments
WAIT FOR ALL TO FINISH
COMBINE all results
TRY UP TO 3 TIMES:
CALL voice cloning service
WAIT 5 seconds BETWEEN TRIES
IF ALL FAIL:
USE default voice
NOTIFY creator about failure
ADD TO QUEUE:
Generate 30 days of content
PROCESS WHEN:
System is not busy
NOTIFY creator when done
Common Use Cases
Copy-paste snippets for frequent operations
CHECK videos exist (3+)
CALL ElevenLabs to clone voice
CALL OpenAI to analyze personality
CREATE ai clone
CONNECT creator to clone as owner
RECORD clone created
GIVE clone ID
CHECK user has payment method
DO TOGETHER charge payment and mint tokens
RECORD purchase
UPDATE balance
GIVE confirmation
CHECK user has tokens
GET conversation history
CALL OpenAI with personality
RECORD interaction
REWARD user with tokens
GIVE response
CHECK course exists
CHECK user not already enrolled
CONNECT user to course as enrolled
RECORD enrollment
GIVE success
GET creator's style
GET top performing topics
CALL OpenAI to generate post
CREATE content
CONNECT creator to content
RECORD published
GIVE post ID
CHECK referrer exists
CHECK new user signed up
CREATE referral connection
REWARD referrer with tokens
NOTIFY both users
GIVE reward amount
The Power of Constraints
By limiting to the ontology's six dimensions, we ensure every feature is type-safe, tested, and composable
For Creators
- ✓ Write what you want in plain English
- ✓ No coding required
- ✓ Can read and understand every feature
- ✓ Can modify features yourself
For AI Agents
- ✓ Unambiguous instructions
- ✓ Maps directly to technical DSL
- ✓ Validates against ontology
- ✓ Generates perfect code every time
For Developers
- ✓ Readable specifications
- ✓ Generated code is clean and tested
- ✓ Can enhance generated code if needed
- ✓ System maintains consistency
The ontology holds the contracts. The DSL speaks those contracts. Together they give humans, AI agents, and code the same, inspectable source of truth.
Command Cheat Sheet
CREATE [type] called [name] WITH [properties]
CONNECT [thing 1] to [thing 2] as [relationship]
RECORD [what happened] BY [who] WITH [details]
GET [what] WHERE [conditions] SAVE AS [name]
CHECK [condition] OTHERWISE [action]
IF [condition] THEN [steps] ELSE [steps]
CALL [service] to [action] WITH [params] SAVE AS [name]
NOTIFY [user] ABOUT [event] VIA [channel]
STREAM [content] TO [platform]
WAIT [duration]
FOR EACH [item] IN [list] DO [steps]
DO TOGETHER: [steps] IF ANY FAIL: [rollback]