> ## Documentation Index
> Fetch the complete documentation index at: https://docs.doconda.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Doconda vs asking ChatGPT or Claude for the file

> ChatGPT and Claude can create files in chat; Doconda creates, reviews and edits them from your app or agent, with a PDF and report on every call.

**Ask ChatGPT or Claude for the file** if it is a one-off document you will create yourself in a conversation and
download by hand. **Use Doconda** if your application or AI agent has to produce the document through an API call, and
you always need the editable file, its PDF and a report of what to check.

## Comparison

| | ChatGPT / Claude (in their apps or with code execution) | Doconda |
| - | - | - |
| What it is | The assistant writes and runs code in an isolated environment and produces the file | API (with Python and TypeScript SDKs and an MCP server) that creates, reviews and edits documents |
| Create | Yes: Word, Excel, PowerPoint and PDF, among others | From a sentence (`prompt`) or your Markdown (`content`): Word, PowerPoint, Excel, PDF and [artifacts](/en/guides/artifacts) |
| Edit existing files | Claude can edit files you upload in the conversation | Yes: you describe the change and the rest stays the same; in Word also with tracked changes and comments |
| Review and repair | Depends on what you ask in the conversation | [Review](/en/guides/review) returns the same file fixed plus a report, without changing your text or values |
| PDF | If you ask for it | Every file comes with its PDF |
| Report on what was delivered | No fixed one | Every document comes with a report: facts not found in your sources, legal citations to check, styles that could not be applied (`style_unsupported`) |
| Templates needed | No | No; you can pass a file to copy its design |
| Use from an AI agent | With the OpenAI API, the Code Interpreter tool generates files in a container that expires after 20 minutes without use | Ready-made function-calling tools for Claude, OpenAI, Gemini and others, plus an [MCP server](/en/agents/mcp) |
| Hosting | On each provider's platform; in Claude, files are downloaded or saved to Google Drive | Hosted service; documents stored in the EU (Ireland) for the [retention](/en/guides/retention) you choose |
| Price | Included in the chat plan (in Claude, it uses more of your usage limit than a normal chat) | Pay per delivered document from a prepaid balance; failed documents are not charged |

## When to ask ChatGPT or Claude

* It is a one-off document and you will create it yourself, in a conversation.
* You already pay for the plan and do not want another account.
* You want to iterate by hand, message by message, until you like the file.
* Downloading the file and checking it yourself is enough.

## When to choose Doconda

* Your application or agent produces the document, not a person in a chat.
* You need a response with the same shape every time: the file, its PDF, page previews and a
  [report](/en/guides/create).
* You want the style applied to the document's styles (Normal, Heading 1…), so anything added later follows the same
  design.
* You have to [edit](/en/guides/edit) your users' files changing only what was asked, or [review](/en/guides/review) them.
* You want to control how long documents are kept and know each document's price in advance.

## Using them together

Your agent built on Claude, ChatGPT or another model decides what to say and Doconda turns it into the file: connect the
tools over [MCP](/en/agents/mcp) (Claude Desktop, Claude Code, Cursor, Codex…) or through the SDK
([Claude](/en/agents/claude), [OpenAI](/en/agents/openai)). You can also pass Doconda the text your model already wrote
(`content`), which is the cheapest way to create.

## Sources

* Claude, create and edit files: [https://support.claude.com/en/articles/12111783-create-and-edit-files-with-claude](https://support.claude.com/en/articles/12111783-create-and-edit-files-with-claude)
* OpenAI, Code Interpreter: [https://developers.openai.com/api/docs/guides/tools-code-interpreter](https://developers.openai.com/api/docs/guides/tools-code-interpreter)
* Doconda: [Create](/en/guides/create), [Edit](/en/guides/edit), [Review](/en/guides/review), [Retention](/en/guides/retention), [Pricing](/en/guides/account#pricing), [MCP server](/en/agents/mcp)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.