How it reads an account
From a handle’s outbound posts and inbound mentions, Xintel computes a full analytics layer. None of it is guessed by a model — it is all derived directly from the gathered data:- Reach and engagement — impressions, likes, reposts, replies, quotes, and bookmarks, each with its own distribution, plus engagement and amplification rates and the best- and worst-performing posts.
- Cadence — whether the account posts in bursts or a steady rhythm, its active hours and weekdays, its peak windows, and its longest silences.
- Composition — the mix of originals, replies, quotes, and reposts; how often posts carry media or links; the language mix.
- Network — edges built from the posts themselves: the accounts this handle most often mentions, quotes, and replies to.
- Topics and information diet — the subjects it returns to and the domains it links out to.
How it models voice
Beyond the metrics, Xintel derives a register — a style sheet of how the account actually writes: its cadence, diction, stance, rhetoric, texture, formatting habits, and the rhetorical devices it leans on. That model is what lets the Post composer draft in a voice that matches the subject instead of a generic one.How it explains and tracks change
Each analysis becomes a report: an AI narrative built strictly on the analytics above — executive summary, strategic assessment, themes, audience read, contradictions, notable posts, and conclusions. Reports are immutable snapshots, so their numbers never drift once written. Because snapshots accumulate, every report after the first carries a change summary — the measured delta from the one before it: new posting volume, metric shifts, topics that are emerging or fading, and accounts entering or leaving the network. Across several reports, a handle becomes a time series rather than a single glance.Data and inference
X data comes through the X API; analysis and media generation run on Venice AI. Xintel does not re-document either — follow those links for their specifics.Who uses it, and what agents get
People work in the app with the full corpus and toolkit — profiles, reports, generation, and dashboards. See Using the app. Agents use the Agent API to buy finished report bodies at $0.01 each over x402. They never receive the profile, posts, or network edges behind a report — that data stays in the app. Humans vs agents covers the split.Names
Xintel is the product. AiSpace is the team that builds it. Venice, X, and Cursor are the tools underneath — dependencies, not co-brands.AiSpace
Xintel is an AiSpace product. The app and these docs share theaispace.bot domain.