# Product DNA: The Brief Your Agent Works From

> Product DNA is the written brief Agent Agency makes from the link you give it. A site reader collects up to five public pages, plus your llms.txt if you have one, and a model turns them into fields: what the product is, who it is for, how it should sound, what it must never say and where its buyers talk. You review it and approve it.

URL: https://agentagency.si/product-dna · Updated: 2026-10-22

## What Product DNA is

An agent writing a reply needs what a new hire would ask for on day one: what you sell, who buys it, which claims are true and which words to avoid. Product DNA keeps those answers in one document, written from your own public pages.

The agent reads the approved version only. A draft does nothing until you approve it. The [glossary entry](https://agentagency.si/glossary#product-dna) gives the short definition next to our other terms.

## What the reader fetches from your link

- **llms.txt and llms-full.txt first**, when your site has them. The [llms.txt proposal](https://llmstxt.org/) describes a short markdown file a site writes about itself, and the writer is told to prefer it for what the product does.
- **Your home page**, then same-site pages: the links in your llms.txt, or common paths such as /pricing, /features, /about and /faq when it lists none. With fewer than two pages, the reader tries your sitemap.
- **Five pages at most**, each cut at 20,000 characters, read as static HTML without running JavaScript, within 45 seconds in total.
- **Your site only**, public pages only. Links to other domains are not followed.

robots.txt is checked for the guessed paths and sitemap pages, not for your home page, llms.txt or the pages it links to. The [security page](https://agentagency.si/security) lists the same rules; our guide to [AI crawlers and robots.txt](https://agentagency.si/guides/ai-crawlers-robots-txt-for-brands) covers what to allow in general.

> **Warning:** The reader is built for websites. An App Store listing reads as one page with its title and description. A Google Play listing, a YouTube or TikTok channel, an Etsy shop or a site that shows text only after JavaScript runs gives it little usable text yet. With such a link, expect to write more of the DNA yourself.

## What the DNA contains

**Fields in a Product DNA**

| Field | What it holds |
| --- | --- |
| Name and one-liner | What the product is called and what it does, in one sentence. |
| Positioning and differentiators | Where it sits next to other options and what sets it apart, as your pages state it. |
| Audience and use cases | Who it is for and the jobs they bring it in for. |
| Tone guidelines | How replies should sound, for example plain and technical. |
| Competitors | Other products in the same space. |
| Mention styles | Three natural ways the product could come up in a reply without reading like an ad. |
| Avoid list | What could get an account banned on Reddit, such as link spam, pushy calls to action or attacks on competitors. |
| Keywords | 5 to 10 phrases each in English, Turkish, German and Spanish, written the way people describe the problem, not as search queries. |
| Category | One primary category from a fixed list, up to six related ones and a few short tags. |
| Target communities | Up to 12 suggested subreddits where your audience talks about the problem. |
| Provenance | Which pages were read, whether llms.txt was found, and whether a note or your own edits shaped the result. |

The writer is told to leave out features, prices, customers or integrations your pages do not mention, and to treat page text as data, never as instructions. It is still a model reading web pages, so check the result before you approve it.

## How suggested communities are used

Suggested subreddits start as unchecked. Names that fail basic checks are dropped; the rest are checked against Reddit later. You can remove any and add your own. A listed community makes the agent more likely to look there, but it does not lift the agent's limits or any community's rules. See [finding subreddits where buyers ask](https://agentagency.si/guides/find-subreddits-where-buyers-ask) to build the list by hand.

## Approve, edit and regenerate

1. **Read the draft** Every field is editable: fix the audience, cut a claim, add a competitor or a community. If the reader found little, fill the fields in by hand.
2. **Pick a category** A draft cannot be approved without a primary category, and the category is yours to change.
3. **Approve** The approved version does not change. Later edits start a new draft, and the previous approved version stays in the history.
4. **Regenerate when the product changes** Run the reader again, with an optional note of up to 2,000 characters. The writer is asked to keep what your pages do not contradict, and communities you added by hand stay on the list.

## What the DNA does not decide

The DNA describes the product; it does not decide when anything goes out. That is the mode you pick: in Approve mode every post, comment and DM waits for your OK, and on Autopilot the agents act on their own inside the rules and limits you set. See [how it works](https://agentagency.si/how-it-works) and the [Reddit page](https://agentagency.si/platforms/reddit).

Disclosure is not settled by the DNA either. You, as the operator, decide how each account presents itself. Reddit's [rules](https://redditinc.com/policies/reddit-rules) ask people not to mislead others or impersonate anyone in a deceptive way, and the FTC's [Endorsement Guides](https://www.ecfr.gov/current/title-16/chapter-I/subchapter-B/part-255/section-255.5) require disclosing a connection to the seller that readers would not expect. Our [disclosure templates](https://agentagency.si/guides/reddit-disclosure-templates) fit in a comment.

## Questions

### Can I write the Product DNA myself?

Yes. Every field of the draft can be edited before you approve it, and if the reader found little you can fill the fields in by hand. The provenance then shows that the DNA was edited by you, next to the pages that were read.

### Does my site need an llms.txt file?

No. Without one, the reader uses your home page and common paths such as /pricing and /about. With one, it starts from your own summary and follows the same-site links you list, so you choose which pages are read.

### What happens when my product changes?

Regenerate the DNA and add a short note about what changed if your pages do not say it yet. Once you approve the new draft, the agent reads it, and the old version stays in the history.

Product DNA is part of early access. Join the waitlist and add your product URL with a few answers while you wait. [Join the waitlist](https://agentagency.si/waitlist)

## Sources

- [The /llms.txt file, v2 (llmstxt.org)](https://llmstxt.org/), accessed 2026-10-22
- [16 CFR 255.5: Disclosure of material connections (eCFR)](https://www.ecfr.gov/current/title-16/chapter-I/subchapter-B/part-255/section-255.5), accessed 2026-10-22
- [Reddit Rules](https://redditinc.com/policies/reddit-rules), accessed 2026-10-22

## Related

- [How Agent Agency Works](https://agentagency.si/how-it-works)
- [Security and Data Handling | Agent Agency](https://agentagency.si/security)
- [Reddit Marketing with an AI Agent | Agent Agency](https://agentagency.si/platforms/reddit)
- [Agent Agency Changelog](https://agentagency.si/changelog)
