Build document collections you can review and trust.
Describe the records you need. Archive AI proposes editable rules and likely matches, while every suggestion stays separate until you include or exclude it.
Collections reflect only documents in your archive. Pocket Archive never assumes a set is complete or includes a suggestion without your approval.
Describe the goal. Review the rule. Approve every inclusion.
Collections add a reviewable organizing layer over your originals. Archive AI can help find likely matches, but the final collection remains yours.
Describe
Describe the records you need
Start with a plain-language goal such as medical expenses for a date range, home warranties, or records for a tax year.
Propose
Review the proposed rule
Archive AI proposes editable topic, document-type, tag, and date criteria, then shows likely matches and dates that need review.
Approve
Include or exclude each match
Review suggestions individually or in bulk. Only included documents are used when you ask a collection question or prepare an export.
Keep the set useful without giving up judgment.
A collection can help you prepare, ask, and export while remaining honest about what is—and is not—in the archive.
- Review likely matchesInspect match evidence, confirm dates, and include or exclude documents one at a time or in a selected group.
Every decision stays visible in the collection’s included, suggested, and excluded views.
- Ask within the approved setCollection questions are scoped to included documents, keeping unreviewed suggestions outside the answer set.
Archive AI keeps supporting documents available so important answers can be checked.
- Prepare a document indexSend included records to the export pipeline and receive an index with links back to the documents.
Exports organize the approved set; they do not replace or silently move the original files.
Only approved records move forward
A living collection with visible decisions.
The system can keep looking for relevant records without hiding how a document entered—or did not enter—the collection.
- RulesEditable rulesReview and revise the topic, document type, tags, and date range that define what the collection is looking for.
- StatesThree clear statesIncluded, suggested, and excluded documents remain separate, so likely matches never look like approved records.
- ReviewNew records stay reviewableWhen newly processed documents match an active collection, they can appear as suggestions for your next review.
Keep collection decisions private and reversible.
Collections can bring sensitive records into one working set. Pocket Archive preserves the originals, keeps suggestions reviewable, and limits collection questions to the documents you approved.
Read the privacy policy- Candidate record
- Encrypted
- Approved set
Does Archive AI automatically include files?
No. Likely matches remain suggestions until you include or exclude them. The proposed collection rule is also editable before confirmation.
How are collection matches selected?
Matches are evaluated against the collection’s topic, document type, tags, dates, and other stored rule criteria. A likely match is still a suggestion, not a final decision.
Can a collection tell me what is missing?
It can show review work within the current archive, but it cannot know about unscanned files, external requirements, or records you have not uploaded.
Which documents does Archive AI use for a collection question?
Collection questions are scoped to included documents. Suggested and excluded records stay outside that answer set.
Build a collection that stays reviewable.
Start free, describe one record set you need, and keep every rule and inclusion decision visible.
Explore document-derived financial insights