Hybrid Reasoning · optional outside help
Outside help. Private details removed.
midas removes names and financial details before asking an outside AI for help. You choose which workflows can use it.
See how it fits into midas .
How it works
Your question. Four steps.
01
Find the facts
midas reads the relevant files in your system.
02
Remove private details
Names and figures become code words. The key stays with you. Uncertain details are removed.
03
Get outside help
Only the redacted request is shared. midas checks and records what goes out.
04
Bring the answer back
midas restores names and figures from your files, checks the answer and includes the sources.
Inside · the document
Candidate: Chan Tai-man
Current salary: HK$62,000 /month
HKID: R123456(7)
Notes: available from October, notice served.
INFORMATION LEAVES THE MACHINE ONCE — IN STEP 03, CARRYING CODE WORDS ONLY
Recorded on the way out · checked against your never-send list · every byte kept in the log
The machine decides when to ask outside. On hard questions, the outside AI brings the outline — your machine, the facts.
The record of what leaves
See what is shared.
Follow an example question from your files to the answer.
Egress ledger
STEP 01 · 03:41:19 HKT
The question, as typed
Priya Raman or Wai-kit Cheung —
who fits Head of Compliance?
Budget HK$1.85M.
Nothing leaves
local · 0 bytes out
1 / 5STEP 02 · 03:41:20 HKT
Found in your own files
2 CVs · the role brief · your files
11 passages · read on the machine
Nothing leaves
local · 0 bytes out
2 / 5STEP 03 · 03:41:22 HKT
What actually left
role: [ROLE_2] · budget: [AMOUNT_7]
candidates: [PERSON_11] · [PERSON_12]
ask: which profile fits, and why
Code words only
1,204 bytes · logged · check: pass
3 / 5STEP 04 · 03:41:24 HKT
The answer, in code words
[PERSON_11] is the stronger fit —
carried [LICENCE_3] through an inspection.
An outline back
inbound · 812 bytes · no identities
4 / 5STEP 05 · 03:41:25 HKT
The answer you see
Priya Raman is the stronger fit —
SFC Type 4 through a first inspection
[CV — Priya Raman.pdf, p.2]
Names put back
3 claims cited · 0 orphans · local
5 / 5The key linking code words to names stays on midas.
Cloud AI asks you to trust a contract. midas asks you to read the log.
Measured results
The results. With the evidence.
0 / 38
Outbound messages that contained anything sensitive
Zero leaks across all testing to date, re-checked independently against a list written in advance.
11 / 15
Canary check — mistakes sent through on purpose that set off the alarm
An alarm that never goes off proves nothing — this one does, and we publish the misses.
13 / 15
Answers with outside help beat the machine working alone
The judges compared answers side by side, not knowing which system wrote which.
12 / 15
Matched or beat handing the outside AI everything, unmasked
Exact figures from your own files, with sources named, beat an outside AI's vague recall. Outside-AI cost per question: about US$0.015.
How we tested
We tested on a set of files invented for testing — no client data — built so that we held a complete list, written in advance, of every name and number that must never leave, so we could search for each one exactly. We recorded what left at the machine's one door to the outside, not in the software's own notes. Answers were compared side by side, and the judges did not know which system wrote which. For details with a fixed pattern — ID numbers, phone numbers, sums of money — the protection is certain by design. Beyond those, it is layered and measured, not absolute, with a physical stop in the path behind both. We do not claim more than that. Certain where certainty is possible; layered and measured where it is not; and a stop built into the hardware behind both.
We didn't trust our own logs until we tapped our own wire.
Bring a question. Follow the answer.
Your IT team can read the record at a seminar, then again on your own machine.