Skip to content
Get started

Documentation

Getting started

titlebase is one program, tbase. It opens a title plant file, searches it, edits it with undo and saves it back, byte for byte where nothing changed. This page takes you from nothing to your first saved edit in a few commands.

What you need

  • A Mac.
  • A title plant file. That is a titlebase file (.titlebase) or any PropertySync plant export (.db). Both open as they are.
  • Enough disk for the plant and a search index beside it (about a fifth of the plant's size).

Get tbase

Builds of tbase come from Limelyte. If you have the source, build the binary with Cargo:

cargo build --release
./target/release/tbase --version

Put tbase on your PATH so the examples below work as written.

Open a plant

Every command takes the plant file as its first argument. Start with file info:

tbase file info sample-county-tx.titlebase
sample-county-tx.titlebase
  size            4.9 MB
  page size       4096
  documents       2400
  landvalidations 4
  autocompletes   11
  rows are        highest id
tbase: file_info in 0.00s

Row counts come from the highest row id, which is instant on any size of plant. Add--exact to count every row. The timing line goes to stderr, so it never mixes with output you pipe.

Find documents

Search by instrument number, instrument type, party and addition. Every filter you give must match.

tbase docs search sample-county-tx.titlebase --type WD --party DOE \
    --addition "SAMPLE ADDITION 5" --limit 3
      5  1994-001288    WD    829/642    1994-11-02  DOE, JANE -> FIRST SAMPLE BANK  [SAMPLE ADDITION 5]
     75  2015-001350    WD    849/610    2015-09-08  DOE, JANE -> ROE, RICHARD  [SAMPLE ADDITION 5]
     80  1991-000492    WD    563/813    1991-03-05  GARCIA, ANA -> DOE, JANE  [SAMPLE ADDITION 5]
3 documents, more after id 80 (--after-id 80)

Pass --after-id to get the next page. Then open one document by row id or UUID to see its full stored JSON:

tbase docs get sample-county-tx.titlebase 75

Without an index, search scans the plant and says so. It stops at --limitmatches, so a search with few matches reads the whole plant. Build the index next.

Auto completes and land validations have their own commands:

tbase ac list sample-county-tx.titlebase --type addition
tbase lv list sample-county-tx.titlebase --addition "SAMPLE ADDITION 5"

Build the index

The index lives in a folder beside the plant, sample-county-tx.titlebase.tbase/. The plant itself is never changed. Build it once; it is reused until the plant changes.

tbase index build sample-county-tx.titlebase
tbase index status sample-county-tx.titlebase

With the index, search adds filed dates, grantor and grantee, lot and block, and book and page, and sorts by any column. --count adds the total.

tbase docs search sample-county-tx.titlebase --addition "SAMPLE ADDITION 5" --lot 4     --sort filed_date --direction desc --limit 3 --count
   142  2026-000982    WD          727/573    2026-07-16  DOE, JANE -> GARCIA, ANA  [SAMPLE ADDITION 5]
   949  2026-001815    MTG        1004/361    2026-05-28  NGUYEN, KIM -> SMITH, ALEX  [SAMPLE ADDITION 5]
  1558  2022-000359    DT          519/010    2022-07-17  DOE, JANE -> SMITH, ALEX  [SAMPLE ADDITION 5]
3 documents
31 matches in all
more: --cursor '{"sort":"filed_date","direction":"desc","key":"2022-07-17","id":1558}'
(indexed: answered from the sidecar index sample-county-tx.titlebase.tbase/index.db)

Sorted results page with --cursor. Party filters can match a prefix or anywhere in the name with --party-match contains.

Edit, undo and save

Edits never touch the plant file. Each write is one journaled run in the same folder beside the plant, and it tells you how to undo it.

tbase docs update sample-county-tx.titlebase 75 --set 'grantees.0.nameFirst=RICHARD A'
updated document 07e58eaf-5e22-04da-6941-01d32e8b3ec8 (row 75): /grantees/0/nameFirst
run 1, op 1
undo: tbase changes undo sample-county-tx.titlebase --run 1

--set takes a dotted path or a JSON Pointer. The value is JSON when it parses and text otherwise. --remove clears a field. Edited rows keep their shape and PropertySync style. Auto completes and land validations have the samecreate, update and delete commands.

tbase changes list sample-county-tx.titlebase      # what changed, per table
tbase undo sample-county-tx.titlebase              # undo the newest run; also tbase redo
tbase changes revert sample-county-tx.titlebase --row 75

When you are happy, save a new file, or save in place and keep a backup:

tbase save sample-county-tx.titlebase --out sample-county-tx-edited.titlebase
tbase save sample-county-tx.titlebase --in-place
saved sample-county-tx-edited.titlebase
  documents       2400 rows: 2399 unchanged, 1 updated, 0 deleted, 0 created
  landvalidations 4 rows: 4 unchanged, 0 updated, 0 deleted, 0 created
  autocompletes   11 rows: 11 unchanged, 0 updated, 0 deleted, 0 created
  verified        true

Save writes a new file and checks it by reopening it. Untouched rows stay byte-identical. To send only what changed, export a change batch: a PropertySync batch with documents, auto completes, land validations and deletions.

tbase changes export sample-county-tx.titlebase --out ./batches

Only one process writes to a plant at a time. A second writer gets a locked error.

JSON output

Add --json to any command. You get one JSON document on stdout and nothing else. Errors exit with status 1 and print an error object. No command ever prompts.

tbase --json file info sample-county-tx.titlebase
{
  "file": "sample-county-tx.titlebase",
  "bytes": 4943872,
  "page_size": 4096,
  "rows": { "documents": 2400, "landvalidations": 4, "autocompletes": 11 },
  "rows_are": "highest_id",
  "extra_tables": []
}
{ "error": { "code": "not_found", "message": "not found: document 99999" } }

Every command's input and output is described by a JSON Schema. tbase commands --json prints them all.

Prove the round trip

check roundtrip parses every row and writes it back with titlebase's own writers, then counts the rows that come back byte-identical. It never modifies the file.

tbase check roundtrip sample-county-tx.titlebase
tbase check roundtrip sample-county-tx.titlebase --copy-to /tmp/copy.titlebase

With --copy-to it also writes a whole new file and compares every row and the schema. On eight real plants, up to 6.3 million documents, every row came back identical.

Connect an agent

Register titlebase with Claude Code, or any MCP client, and your agent gets every command as a tool:

claude mcp add titlebase -- tbase mcp

See MCP server for other clients and for binding the server to one plant.