canonical: https://jentic.com/apis/finestructure.ai/finestructure

# Fine Structure API

The Fine Structure API builds and manages complete web applications from natural-language prompts. It creates and updates apps, reads and writes files in each app's virtual filesystem, defines and queries data entities, and stages file changes it can validate before applying. It publishes apps and manages their subdomains, custom domains, SSL, and redirects, runs backend functions behind access policies, and operates autonomous agents with scheduled tasks on your account. Generation work is asynchronous, so jobs are created and then polled for status.

## For AI agents

Build and update web apps from prompts, edit their files and data entities, validate and apply change sets, publish and route custom domains, run backend functions, and operate autonomous agents.

## Scope

Does not handle hosting of apps built outside the platform, standalone source-control, or general CI pipelines for external repos. Use for building, editing, and publishing Fine Structure web applications only.

## Capabilities

- Create a web application from a natural-language prompt and update it with new prompts
- Read, write, patch, and rename files in an app's virtual filesystem
- Define and query data entities and create, update, or delete their records
- Publish an app and manage its subdomain, custom domain, SSL, and redirects
- Run an app's backend functions and set route, function, and entity access policies
- Create autonomous agents and schedule their tasks on your account

## Use cases

### Agent-built web apps through Jentic

An AI agent turns a brief into a working app: it calls create_app with a prompt, polls the generation job until it is done, reads the file tree, patches a page, then publishes. Because file reads, writes, and validation are separate operations, the agent can inspect and fix its own output before anything goes live.

Example prompt: Create an app from a prompt with create_app, poll get_job_status until done, then publish the app

### Manage app data and backend logic

An app is more than its pages. Fine Structure exposes each app's data entities and lets an agent query records, create and update them, and run backend functions behind route and function policies, so an integration can seed starter data, wire up logic, and keep access rules server-owned rather than guessed in the frontend.

Example prompt: Query records from an entity, create a new record, and invoke a backend function to verify it responds

### Publish and route a custom domain

Shipping an app means getting a domain live with SSL. The API publishes an app, sets its subdomain, attaches a custom domain, returns the DNS records to verify it, and reports SSL status, so a deployment workflow can take an app from generated to reachable on its own domain without the dashboard.

Example prompt: Publish an app, add a custom domain, and read the DNS and SSL verification status it returns

## Key endpoints

| Method | Path | Description |
| --- | --- | --- |
| POST | `/api/mcp/tools/create_app` | Create an app from a natural-language prompt |
| POST | `/api/mcp/tools/update_app` | Update an app with a new prompt |
| POST | `/api/mcp/tools/publish_app` | Publish an app and make it live at a URL |
| POST | `/api/mcp/tools/write_app_file` | Create or replace a file in an app |
| POST | `/api/mcp/tools/query_entity` | Query records for one app entity |
| POST | `/api/mcp/tools/get_job_status` | Poll an async generation job |
| POST | `/api/mcp/tools/create_agent` | Create an autonomous agent on the account |
| GET | `/api/mcp/manifest` | Tool discovery manifest with names and schemas |

## Key resources

- **Apps** — Create, update, publish apps and read their files, pages, entities, and status
- **Entities** — Define data schemas and query, create, update, and delete records
- **Domains** — Manage subdomains, custom domains, SSL, and redirect policy
- **Agents** — Create autonomous agents and schedule their tasks

## AI readiness

This API is usable in Jentic One now. Its AI-readiness score against Jentic's framework shows where it stands today and where improvements would make it even easier for agents to use.

- **Score:** 64 / 100
- **Maturity:** AI-Aware
- **Dimensions:**
  - Foundational Compliance: 73 / 100
  - Developer Experience & Jentic Compatibility: 68 / 100
  - AI-Readiness & Agent Experience: 50 / 100
  - Agent Usability: 75 / 100
  - Security: 60 / 100
  - AI Discoverability: 76 / 100
- **View full report:** https://jentic.com/apis/finestructure.ai/finestructure/scorecard
- **How the score is calculated:** https://docs.jentic.com/reference/api-readiness-framework/overview/
- **More about the dimensions:** https://docs.jentic.com/reference/api-readiness-framework/specification/#dimensional-model-overview

### Score it yourself

Every API in the directory is allowlisted, so you can re-score it with no key required.

- **Score your own API:** https://jentic.com/scorecard.md
- **Scoring CLI agent skill:** https://github.com/jentic/jentic-api-scorecard/blob/main/skills/jentic-api-scorecard/SKILL.md

```sh
npx @jentic/api-scorecard-cli score <openapi-url>
```

## Why Jentic

- **Setup:** Wiring the Fine Structure API by hand means handling its Bearer token, the asynchronous job polling for every generation, and the stage-validate-apply flow for file changes yourself. Through Jentic you install once, import it from the API Directory, store the token once, and your agent calls it.
- **Permission scoping:** App ids travel in the request body, not the URL path, so rules bound which operations your agent may call rather than which app. You choose the operations it may call, so destructive ones like delete_entity_records, remove_custom_domain, or delete_secret stay out unless you add them.
- **Credential handling:** Your Fine Structure token is stored once, encrypted, by your own Jentic One instance and injected at execution time. It never enters the agent's prompt, logs, or context.
- **Discovery method:** Agents search Jentic by intent such as 'build a web app from a prompt' or 'query an app entity', and Jentic returns the matching Fine Structure operation with its input schema so the agent calls the right endpoint without browsing the reference docs.

## Related APIs

- **Webflow** — Webflow builds sites and apps visually, an alternative to Fine Structure's prompt-driven generation.
- **Backendless** — Backendless is a no-code backend and app platform, an alternative to Fine Structure's full-stack generation.
- **Vercel** — Vercel deploys and hosts frontends that a generated app could target.

## FAQ

### What authentication does the Fine Structure API use?

It uses a Bearer token sent in the Authorization header, per its OpenAPI spec. You create a token on your account and the same token scopes calls to the apps and agents you own, while a tool manifest and a health check are open.

### Can I manage an app's data and files with the Fine Structure API?

Yes. You can read the file tree, read and write individual source files, and patch or rename them, and separately define data entities and query, create, update, or delete their records. Changes can be staged in a change set, validated, and applied, so an agent can check its edits before they land.

### What are the rate limits for the Fine Structure API?

The OpenAPI spec does not state a request rate limit. Generation work is asynchronous and metered by your account plan, so create a job and poll its status rather than busy-waiting, and check the Fine Structure documentation for current limits.

### Is there a Fine Structure MCP server?

Yes, Fine Structure exposes an MCP endpoint and a tool manifest. You can also connect Fine Structure directly through Jentic: your agent gets the full 130-endpoint surface without loading another server's tool definitions into its context, your token is injected at call time by your self-hosted instance instead of living in a config file, and your rules decide which operations the agent may call.

### Can I limit what my agent is allowed to do with the Fine Structure API?

Yes. Write a rule that allows the read and edit operations your agent needs, such as reading files and querying entities, and leave the record-delete, secret, and domain-removal operations off the list, so an agent can build and inspect an app without removing data or rotating secrets. Every call it makes is recorded in your Jentic One audit log.

### How do I build an app with the Fine Structure API through Jentic?

Search Jentic for 'build a web app from a prompt', add the Fine Structure API from the directory, and store your token once; your agent then creates the app, polls the generation job, edits files, and publishes. To run it on your own infrastructure, install Jentic One from its GitHub repo.