aichivist
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| get_rulesB | The person's filing rules. Read first. |
| surveyC | Stock-take of folders: sizes, types, duplicates, unclear names, stale files, drive health. |
| list_folderC | Entries of a folder (first 100; 'total' gives the count). |
| read_documentB | Text of a document, first page by default (scans via local OCR). next_offset continues. |
| suggest_namesC | For all files in folders: date, issuer, kind and a suggested name, one row each. Start here. |
| inventoryC | Files in a folder; for areas also what changed or was moved by hand since the last run. |
| find_duplicatesC | Byte-identical files: new items vs. a target area and among themselves. |
| find_similarC | Near-duplicates by text (new scan or re-download of the same document). |
| suggest_nameC | Date, issuer, kind, summary, keywords and a name for one document. confident=false: read it. |
| naming_examplesC | How the person named files before – match this style. |
| structure_adviceC | Structure problems of an area (too deep, empty, by type, duplicate names) to discuss. |
| retention_dueC | Records older than a retention period – ask per record: archive or delete. |
| plan_showD | Rows of a filing plan. |
| plan_checkB | Simulate the plan (missing folders, closed collections, clashes, program data). Changes nothing. |
| program_checkB | Does a program depend on these paths (program folders, libraries, links, open files)? |
| storage_checkC | 3-2-1 backup check and yearly cost from a storage table. Never invent prices. |
| statusD | Is a folder an open or a closed collection? |
| verifyC | Check a closed collection against its fixity manifest. |
| searchC | Full-text search in a closed collection. |
| plan_addC | Add one plan row: source, target path (or TRASH) and reason. |
| plan_photosB | Plan rows for photos/videos: <photos_root>//, year from the capture date. |
| plan_reviewC | Open the plan in the person's browser to tick and approve rows. |
| catalog_checkC | Is every file of an area in its catalog? add_missing appends rows to describe. |
| catalog_describeB | Fill one '⚠ to describe' catalog row: content, search terms, date from the document. |
| plan_approveB | Approve plan rows ONLY after the person said yes in chat. leave_link only if they chose it. |
| plan_applyA | Execute approved rows (logged, undoable). create_folders only if the person approved them. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
| tidy_up | File the documents in a folder with a plan you approve |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 26 tools
Most tools target clearly different stages of filing work, and descriptions distinguish the close ones reasonably well. A few clusters remain easy to confuse: list_folder/inventory/survey and suggest_names/suggest_name/naming_examples require careful reading before selection.
All names use snake_case and several useful prefixes (plan_, catalog_) group related tools. However, the set mixes verb_noun, noun_noun, and bare verb/noun conventions, so there is no single predictable pattern.
At 26 tools, the surface is heavy for a personal filing assistant and many operations could likely be consolidated. The domain is broad, but the tool count exceeds a comfortably scoped set.
The set covers the main lifecycle: inventory, naming suggestions, duplicate detection, planning, approval, application, verification, search, and catalog maintenance. Minor gaps exist around direct search/update operations outside closed collections and direct editing of rules, but core workflows are covered.