Structure before generation
The language parser maps code relationships, giving the agent a view of how your project fits together before it starts writing.
BUILD WITH LESS FRICTION
Bring your code
We'll handle the docs
Deltify is an agentic AI documentation platform. Deltify looks at your entire code, reads it and converts it into readable interactive documentation.
Join the waitlist for early accessSee how it works
YOUR REPOSITORY
acme / orbitmain
src/
routes/projects.ts
services/project.ts
types/project.ts
README.md
routes/projects.ts
router.post('/projects', async (req) => {
const project = await createProject({
name: req.body.name,
});
return project;
});Create a project
Create a new project in your workspace.
POST
/projectsRequest body
namestring · requiredThe name of your new project.
{ "name": "My first project" }JSON Start with your sourceA connected repository gives Deltify the code to understand.
HOW THE AI GENERATES YOUR DOCS
The agent writes. Context comes first.
Connect your repository and give it direction. Our language parser and internal system prepare focused project context, then an AI agent uses a large language model (LLM) to generate your documentation.
- 01
Connect your repository
Bring your GitHub repository. Your code is the starting point.
- 02
Give it some direction
Choose your audience. Add instructions, images, or the details only you know.
- 03
Build the project context
The language parser maps code structure and relationships. Our internal system prepares relevant context for the LLM.
- 04
Let the AI agent generate
The agent uses the LLM to turn prepared context into guides, references, and examples for your audience.
PROJECT CONTEXT, BEFORE THE FIRST WORD
Context first. Generation second.
Our language parser maps the structure and relationships in your repository. Deltify's internal context system combines those results with relevant code and your instructions, then passes focused input to the LLM. The AI agent uses that context to write about your project and its behavior.
Built to reduce generic output and unsupported assumptions.01
Context the LLM can use
Our internal system combines relevant parser output with your instructions and audience. The model receives focused input for the documentation it's generating.
Specific to your project
Source context gives the agent concrete details to explain behavior and write examples. This is designed to reduce generic answers and unsupported guesses.
HOW PROJECT CONTEXT REACHES THE MODEL
Your repository
Code + your instructions
Language parser
Structure + relationships
Context system
Focused input for the model
AI agent + LLM
Generation for your audience
Documentation
Guides, references + examples
BEYOND STATIC MARKDOWN
Read it. Try it. Build with it.
Test an API, see a component in action, and follow a guide with the context right there. Give readers a place to explore and build with your project.
DOCUMENTATION THAT KEEPS UP
Push your changes. Your docs follow.
Every push to your connected GitHub repository triggers a documentation update. Deltify refreshes the project context, and the AI agent updates the docs to reflect your code. You keep shipping; the docs keep up.
A push starts the update
Changes reach your connected GitHub repository.
Fresh code, fresh context
The language parser and internal system refresh the context for the LLM.
The AI agent takes it from here
The agent uses the updated context to generate documentation that reflects the changes.
Your documentation catches up
New code is explained, without a manual docs rewrite.
One less thing on your release checklist.
Add project descriptionsmain
types/project.ts
type Project = {name: string;description?: string;};
New commit received
API referenceUpdate pending
Project
namestring · required- The name of your project.
- description
- A new field is on its way.
ONE CODEBASE. DIFFERENT READERS.
Write for the person on the other side.
An API reference and a getting-started guide have different jobs. Give each audience the explanation they need, from a first walkthrough to an integration guide.
Frequently asked questions
Some of our most asked questions
What is Deltify?
Deltify is an agentic AI documentation platform. Connect a GitHub repository and give the agent your instructions. Our language parser and internal context system prepare project context, and the agent uses an LLM to generate documentation for your audience.
How is this different from asking my agent to write documentation?
Deltify automates this with a few clicks. Furthermore Deltify has an advanced internal context system specifically optimized for this. Deltify also isnt just a static documentation generator, it features advanced interactive api testing playgrounds, chatbots, component rendering and more.
Is this only for developer documentation?
No. Create approachable guides for everyday users, API references for developers, and codebase overviews for teammates. Your instructions help the agent write for the audience you've chosen.
What can readers do inside the docs?
Test API requests in a playground, explore rendered component examples, and follow guides with code and images. Readers can try things alongside the explanation, with the context they need in one place.
Do the docs update when I push changes?
Yes. Each push to your connected GitHub repository triggers an update: our language parser and internal system refresh the project context, then the AI agent updates the documentation to reflect the new code. You won't need to rewrite the docs by hand after every change.
When can I try it?
Join the waitlist to hear when early access is ready.
BUILD WITH LESS FRICTION.
Let your code do the explaining.
Join the waitlist for Deltify. The Ai powered documentation platform that understands your project and keeps up with your code.