Sources
PDFs, text files, pasted text, a web page fetched once, and workspace items you may see.
Answers you can check against the source
Collect sources in a notebook, ask questions answered only from them with citations, and write a brief.
Notebooks follow the workspace people model: private to the author until shared with chosen team members or the whole team to read and ask or to edit; owners and the admin never open a notebook not shared with them. Sources are PDFs up to 15 MB in R2, .txt and .md files, pasted text, a web page fetched once on request (private and internal addresses refused; the readable text kept with the URL and fetch date) and workspace items found through Workspace Search as the person and copied, with a Refresh; up to 50 per notebook. A question uses up to 20 selected sources, sent to Claude (claude-opus-5-5) as document and search-result blocks with citations on; the answer streams to the page, each citation opens the cited range or PDF page, and an answer without a citation says it could not find it in the sources. Questions are private to who asked. Each source gets a guide (summary, key topics, questions to ask). Answers and quotes are pinned as findings in sections with notes beside them; a research brief (summary, findings, evidence, open questions, recommendations) is laid out from the findings with numbered source notes, or drafted by Claude from them, edited with version checks, shared read-only as /r/<token>, downloaded as Markdown, printed to PDF or saved to Writer as a private document. Owners set AI requests per person per hour and tokens per person per month, and see the usage. Webhooks notebook.created, notebook.shared and brief.published, and a workspace feed of notebooks and briefs.
What Research does for you.
PDFs, text files, pasted text, a web page fetched once, and workspace items you may see.
Each answer cites the passage or PDF page it came from, and says so when the sources do not answer.
Pin answers and quotes into sections, then write a brief to share as a link, print or save to Writer.
A suggested path for Research. What you can do depends on your access in the app.
It is private to you until you share it with people or the team.
Upload, paste or add a web page, choose the sources a question uses, and ask.
Lay out the findings by section, draft the rest with AI if you want, and publish it.
Live host, inventory, governance, health checks, verification, dependencies and history.
Notebooks follow the workspace people model: private to the author until shared with chosen team members or the whole team to read and ask or to edit; owners and the admin never open a notebook not shared with them. Sources are PDFs up to 15 MB in R2, .txt and .md files, pasted text, a web page fetched once on request (private and internal addresses refused; the readable text kept with the URL and fetch date) and workspace items found through Workspace Search as the person and copied, with a Refresh; up to 50 per notebook. A question uses up to 20 selected sources, sent to Claude (claude-opus-5-5) as document and search-result blocks with citations on; the answer streams to the page, each citation opens the cited range or PDF page, and an answer without a citation says it could not find it in the sources. Questions are private to who asked. Each source gets a guide (summary, key topics, questions to ask). Answers and quotes are pinned as findings in sections with notes beside them; a research brief (summary, findings, evidence, open questions, recommendations) is laid out from the findings with numbered source notes, or drafted by Claude from them, edited with version checks, shared read-only as /r/<token>, downloaded as Markdown, printed to PDF or saved to Writer as a private document. Owners set AI requests per person per hour and tokens per person per month, and see the usage. Webhooks notebook.created, notebook.shared and brief.published, and a workspace feed of notebooks and briefs.
Quick questions over the whole workspace belong to Assistant; finished documents belong to Writer.
5 of 18 required controls are configured. Catalog deployment and inventory maturity remain independent until every required control is present.
researches.appneural.com
No remediation required.
No remediation required.
No remediation required.
No remediation required.
Configure or repair: Repository exists.
Configure or repair: Business owner exists.
Configure or repair: Product owner exists.
Configure or repair: Technical owner exists.
Configure or repair: Maintaining team exists.
Configure or repair: Documentation exists.
Configure or repair: Architecture documentation exists.
Configure or repair: Security review is current.
No remediation required.
Configure or repair: Dependencies exist.
Configure or repair: Consumers exist.
No remediation required.
Configure or repair: SLA or SLO exists.
Configure or repair: On-call exists.
Configure or repair: Release exists.
No remediation required.
Configure or repair: Infrastructure drift is clear.
No remediation required.
No remediation required.
No remediation required.
No remediation required.
Configure or repair: Repository exists.
Configure or repair: Business owner exists.
Configure or repair: Product owner exists.
Configure or repair: Technical owner exists.
Configure or repair: Maintaining team exists.
AI outputs are saved as drafts with evidence references. They never replace verified catalog fields automatically.