Attribution and Content Sourcing Policy
Effective: Conflab v0.1.12+
This document sets out Conflab's position on crawling, licensing, and credit for catalog content that originates outside of Conflab. It covers what we collect, how we credit it, how we respect upstream licences, and how to request removal.
Scope
This policy applies to every Entry in the Conflab catalog that was imported from an external source rather than authored directly on the platform. It does not apply to:
- Entries authored by Conflab users through the publish flow. Those are governed by the Terms of Service and the UGC Safety Policy.
- Seed content authored by the Conflab team. That content carries its own licence declaration.
When in doubt, the presence of a non-empty source_url on an entry indicates that this policy applies.
For the licence you grant to Conflab when you publish your own content (including the AI-training rights clause), see Terms of Service §5.3 and §5.4.
What we collect
For every imported entry, Conflab captures and persists:
- Original author name -- the upstream author's byline as it appears on the source, or an identifying handle if no byline is available.
- Source URL -- the canonical upstream URL for the entry. This is the link users follow to reach the original.
- Source name -- a human-readable identifier for the source site or collection (eg "Awesome ChatGPT Prompts").
- Source licence -- the licence under which the upstream content is distributed, captured as an SPDX identifier where known (eg
MIT,CC-BY-4.0,Apache-2.0), or as free-text when the upstream uses a non-standard declaration. - Fetched-at timestamp -- the UTC date and time at which our crawler retrieved the entry.
These five facts are persisted as distinct attributes on the Entry resource. They are not collapsed into a single free-text field, and they are not derived or inferred from other data.
How we credit
Every imported entry renders an attribution block on its detail page. The block names the upstream author, links the source URL, names the source, and states the licence we observed. If any of these facts are missing for a given entry, the block renders the parts we have -- it does not fabricate.
The attribution block sits immediately below the Conflab fork-attribution block and above the rating widget, so provenance is visible before social signals.
This policy is linked from the site footer alongside Privacy, Terms, About, and UGC Safety.
How we respect upstream licences
Our crawler classifies the upstream licence of every candidate entry into one of four buckets. The bucket determines what happens next:
| Bucket | Examples | Action |
|---|---|---|
permissive |
MIT, BSD, Apache-2.0, CC0, CC-BY-*, public domain | Imported with attribution. No further restrictions flagged. |
conditional |
CC-BY-SA, CC-BY-NC, GPL family | Imported with attribution, and the licence terms are recorded verbatim so re-licensing constraints are preserved. Sources that disallow commercial reuse are respected accordingly. |
prohibited |
Proprietary, no licence declared, explicit "no redistribution" | Not imported. The source is skipped and logged. |
uncertain |
Source exists but the licence cannot be determined with confidence | Flagged for human review before any import decision is made. |
Classification is advisory and pattern-based. It is not a substitute for legal advice, and we do not claim it is. Where a source's licence is genuinely ambiguous, we default to the more restrictive reading.
Robots.txt and rate limits
Our crawler respects robots.txt for every source it visits. It enforces per-source rate limits configured in the source registry, with a conservative default. Sources are crawled on a schedule, not continuously.
If a source's terms of service disallow automated retrieval, we do not crawl it -- regardless of what robots.txt permits.
Removal on request
If you are an upstream author (or an authorised rights-holder) and you want content removed from the Conflab catalog, or you believe an entry is misattributed, contact us at conflab-crawler@geodica.com with:
- The URL of the Conflab entry.
- Your relationship to the upstream source.
- The action you want taken (remove entry, correct attribution, update licence record).
We aim to acknowledge within five business days and to act within seven business days from acknowledgement. Removals are logged; we retain a record of the request and the action taken so that future re-crawls do not re-introduce removed content.
We do not require formal legal notices for removal requests, but we do verify the requester's connection to the upstream source before acting.
Re-crawling and updates
Crawls are periodic. A re-crawl of an already-imported entry refreshes the fetched_at timestamp and updates the source licence record if the upstream declaration has changed. It does not silently overwrite curator edits made on the Conflab side -- those are preserved.
If a previously-imported source becomes inaccessible, unreachable, or explicitly prohibits reuse, affected entries are flagged for review. We retain them pending review rather than deleting automatically.
What this policy does not do
- It does not replace the upstream licence. If you use an imported entry, you are bound by the upstream licence terms, not just this policy.
- It does not claim copyright on imported content. Conflab is a curator, not the copyright holder.
- It does not cover content moderation. See UGC Safety for moderation posture on user-authored content.
Contact
Questions about this policy: conflab@geodica.com.