Funder-ready numbers from the records you already keep.

Volunteers keep working as they do. Every number links to the row, message or photo it came from.

Made for small charities with no data staff.

View a sample report

Made-up records. No account needed.

A made-up charity, October 2026

Referrals, G.T.

…This is to confirm that we received your referral of 31 October for Mr Senthil Raj (high blood pressure). Mr Senthil Raj has been registered with our outpatient clinic and was given an…

Recorded asReferralP-0056S.R.Health Outreach

Your team sees names. Your funder never does.

13

Referrals made

Times someone was sent on to another organisation or service.

Showing 1 of 13 records.

See how a record is counted

Follow one record

One person at English Classes in October 2026, from the register to the report.

English Class Attendance
AAM
1Name25/10/2026
26Lokman✓

A tick in the register

One person's row in English Class Attendance, with 4 ticks in October 2026.

Type
Attendance
Person
P-0010L.
Programme
English Classes
Date
25 Oct 2026
How sure
Very sure
Status
Counted

English Class Attendance, Attendance

Name: Lokman Date: 25/10/2026 Attended: ✓

The record

Each tick becomes one record. The person appears as a code and initials.

Oct 2026 figures

People served

97

Up 17% compared with Sep 2026

from 139 rows, messages and files

Open the records behind this figure

Showing this person's 4 of its 197 records.

  • English Class Attendance, Attendance

    Name: Lokman Date: 04/10/2026 Attended: ✓

    Recorded asAttendanceP-0010L.English ClassesCounted in People served

  • English Class Attendance, Attendance

    Name: Lokman Date: 11/10/2026 Attended: ✓

    Recorded asAttendanceP-0010L.English ClassesCounted in People served

  • English Class Attendance, Attendance

    Name: Lokman Date: 18/10/2026 Attended: ✓

    Recorded asAttendanceP-0010L.English ClassesCounted in People served

  • English Class Attendance, Attendance

    Name: Lokman Date: 25/10/2026 Attended: ✓

    Recorded asAttendanceP-0010L.English ClassesCounted in People served

4 classes, one person

What these classes add in October 2026

What these classes add in October 2026
FigureAdds
People served+1
Attendances+4
Times helped+4
The same class in a chat

A chat message about the same class would join this record as a second source. It would still count once.

What if no names were written down?

HLWC Programmes

Title: Health screening - Jurong dorm When: Sun 11 Oct 2026, 09:00 to 13:00 Where: Jurong dorm Guests: 4 Description: Blood pressure, glucose and eye checks. Referral letters for the polyclinic. Update: 47 workers screened, 5 referred to the polyclinic.

A group with no names, found in 4 places and counted once.

A group with no names, found in 4 places and counted once.
FigureAdds
People served0
Attendances+47
Times helped+1

Harbour Light Workers Centre

By programme

English Classes

Figures for October 2026
FigureThis periodPeriod before
People served2017
Times helped5136

October 2026

In the report

This person is one of 20 people served at English Classes in October 2026.

A report you can send your funder

  • Your figures for the period, with a chart for each programme.
  • A last page on how each figure was counted.
  • Your own words, and photos you have consent to share. Never names or records.

Harbour Light Workers Centre

Impact report

October 2026

Key numbers

People served
97
Up 17% compared with September 2026
Times helped
202
Up 22% compared with September 2026
Websites built
15
Up 15% compared with September 2026

The period in numbers

Figures for October 2026
FigureThis periodPeriod before
People served9783
Websites built1513
Stories written1310
Times helped202166
A made-up charity's report for October 2026

Sources

See every source

You check what it isn't sure of

A record the AI isn't sure of waits for you. It isn't counted until you answer.

Privacy

Your funder sees totals, never your records or the names in them.

  • Evidently reads only what you pick. It never posts, sends or changes anything.
  • Claude, an AI model made by Anthropic, does the reading. In text, phone and ID numbers, emails and addresses are hidden from it first. Photos are sent as they are. It sees names, because it needs them to tell people apart.
  • Delete a source and everything it brought in goes. Delete a person and their records go. Messages that name them stay until you delete their source. The numbers are counted again.

Using Evidently

Who can use it
We're piloting with a few invited groups. You can sign up and try it, or ask to join the pilot. The sample is open to everyone.
Cost
Free during the pilot. Pricing after the pilot isn't set yet.
Made by
OpenSkills, a pro bono studio
Your data
Evidently is new. It has no security certification or independent audit yet. Data protection

Try it

Open a made-up charity with 10 sources and records from February to November 2026.

Sign up