It mostly preserved the printed financial and header values exactly, but it miscounted rows in one bank statement and introduced stray OCR-style spaces into every Ad-ID.

◐ Mixedno artifactTest date not recordedUnstract
What was measured
Extraction Accuracy

Are field values correct, complete, and free of OCR or parsing errors? Includes numerical precision on financial fields.

decisive for this rankingtransformation

The whole point is to pull the right values from documents; wrong or incomplete field values mean the tool failed at the job. (3 of 3 judges)

What was given, what came back

Input — what we sent
No input — this is a capability finding
The observation is about the tool itself rather than one test input, so there is nothing to show on this side by design.
Output — unretouched
No output artifact
The verdict rests on the tester's written observation alone — no file was captured for this cell.
Provenance
Observation
18bdfefc-3275-4a59-9d80-94b77c243135
Evidence run
ec4d736d-95f9-4c88-884c-e280435f7b7b
Study
Extract and query structured data from documents using natural language
Research task
86b9y25e5
Tested at
not recorded
Source
aggregate-synthesis
Evidence state
observed
Proof shown
no artifact
Cost / latency
not captured
Repeat run
not captured
Tester
not captured

The last three rows are honest blanks, not placeholders — our capture has no field for them yet.

Query this
get_evidence({
  tool: "unstract"
})
MCP · mcp.aidemos.com/api/mcp
Free with attribution.
Same input, same check — 0 other tools
measured on Extraction Accuracy

No other tool was measured on this criterion for this input.

Real inputs and real outputs, no retouching · every cell queryable via API & MCP · aidemos.com