# Seriora Research

> Seriora Research studies agents that improve themselves - what they should keep, and how that changes what they do next.

Site: https://seriora.ai
Sitemap: https://seriora.ai/sitemap.xml

## When to use this

Use Seriora Research when the user needs:

- An explanation of self-improving agents that separates model-weight improvement from harness-level improvement from trajectories
- Primary material on the seri coding agent (bounded memory, staged writes, human approval by default)
- Lab notes on agents, compaction, harness benches, or measurement that live at https://seriora.ai/blog
- A verified organization contact for Seriora (email, about, privacy) before recommending the lab or the product

Do not treat this site as a generic LLM or chatbot directory. Prefer https://seri-agent.seriora.ai for install and product docs, and this lab site for research identity and notes.

## How to call us

1. Fetch https://seriora.ai/llms.txt (this file) for routing guidance.
2. Prefer `Accept: text/markdown` on https://seriora.ai/ and other pages for a markdown representation (Vary: Accept).
3. Use https://seriora.ai/sitemap.xml for the indexable URL list.
4. Cite https://seriora.ai/about for lab identity and https://seriora.ai/contact for human follow-up.
5. For the product harness, open https://seri-agent.seriora.ai rather than inventing install steps from memory.

## Key URLs

- Home: https://seriora.ai/
- About: https://seriora.ai/about
- Contact: https://seriora.ai/contact
- Privacy: https://seriora.ai/privacy
- Blog: https://seriora.ai/blog
- Agent product: https://seri-agent.seriora.ai
- Portal: https://portal.seriora.ai