Research is only useful if other people can read it, verify it, and build on it. Webhound now lets you publish any session or folder as a versioned, citable artifact with a permanent URL. Anyone can browse it. Anyone with the right license can fork it into their own workspace and keep going where you left off.
This isn't a PDF export. A published Webhound session is a live document — with inline citations, source lists, claim traces, datasets, and the full research trail. Readers can click through every source, inspect every claim, and see exactly what the agent did to produce the result.
What gets published
When you publish, the system takes a full snapshot of everything in the session or folder at that moment:
- •Reports — the output document, working documents, all inline citations and claim traces.
- •Datasets — every row, the schema, enrichment results.
- •Sources — a deduplicated list of every URL the agent visited, extracted from citations, the knowledge base, and dataset rows.
- •Verification results — if the session was audited, the full claim-by-claim verification data comes along.
- •Conversation history (optional) — the full message trace between you and the agent, so readers can see the prompt, the plan, and every course correction.
The snapshot is immutable. Your workspace keeps evolving — run new sessions, edit folders, add data — but the published version stays exactly as it was. Readers always see a frozen, reliable state.
Versioning
Every time you republish, a new version is created. Old versions are never overwritten — they stay accessible at /p/your-slug/v/1, /v/2, and so on. The main URL always points to the latest version.
After version 1, every update requires a note explaining what changed. The History tab shows the full version timeline — every publish date, every update note, and if the publication is a fork, where it branched from.
Licenses
You choose a license when you publish. There are four levels, and the system enforces them:
- •View Only. Anyone can read it. Nobody can copy it into their workspace.
- •Attribution. Readers can fork it and use it in their own work, but they can't republish the copy as a new publication. Attribution back to you is preserved.
- •Open. Fork, modify, republish. Derivatives must use the same license or stricter. Full lineage is tracked.
- •Open + Commercial. Same as Open, but explicitly allows commercial use.
The license isn't just a label. The copy button disappears for View Only publications. Attribution and Open publications that get forked carry the license forward into the derivative — the system won't let someone republish under a more permissive license than the original.
Forking
If the license allows it, any logged-in user can fork a publication into their own workspace with one click. The fork creates a full copy — sessions, documents, datasets, sources, and optionally the conversation history. Everything lands in a new folder in your workspace, ready to extend.
Lineage is tracked automatically. If you republish a fork, the new publication links back to the original. The History tab on both the original and the derivative shows the relationship. The Attribution tab lists every researcher who contributed to the chain.
This is the core idea: research compounds. Someone publishes a $50 report on market dynamics. You fork it, run three more sessions on a specific vertical, and publish your extended version. A third person forks yours and adds a dataset of companies. Each step builds on real, traceable prior work — not a prompt someone typed from scratch.
Explore and discovery
The /explore page is the public index. It lists every public publication with the title, author, source count, cost, and date. Search works across titles, summaries, and author names.
You can follow authors and watch specific publications. When someone you follow publishes something new, or a publication you're watching gets an update, it appears in your feed. Notifications are generated for both.
Author profiles
Every user who publishes gets an author profile at /author/{id}. It shows your display name, bio, expertise areas, external links, and a list of all your publications with aggregate stats — total copies, total sources, total spend.
Profiles are public. Other users can follow you from your profile page or from any of your publications. Follower and following counts are visible on the Explore page's network tabs.
The publication page
Each publication lives at /p/{slug}. The reader sees:
- •Documents tab — the full report or dataset, rendered exactly as it appears in the workspace. Sidebar navigation for multi-session folders. Click any report to read it with inline citations and claim trace dots.
- •Sources tab — every URL the agent visited, searchable and filterable by document.
- •Activity tab — the publish/update/fork timeline. Shows who published, when, and links to any derivatives.
- •Attribution tab — the full chain of contributors. Who authored it, what it builds on, what cites it, and what's been forked from it.
- •History tab — immutable version timeline with update notes and fork origin.
API
Everything above is available programmatically. The API v2 endpoints cover the full lifecycle:
- •POST /publications — publish a folder or session.
- •GET /publications — list and search public publications.
- •GET /publications/:id — full publication with current version snapshot.
- •GET /publications/:id/versions — version history and fork list.
- •POST /publications/:id/copy — fork into your workspace.
- •POST /publications/:id/watch — subscribe to updates.
- •GET /publications/:id/attribution — the full contribution and lineage graph.
Questions? Email team@webhound.ai.