Publish without hedging
Conservatism is in the math: a 50% discount on raw signal and a 0.85 confidence cap. The scores you put in front of funders, boards, or regulators are bounded by mechanism — not promised by marketing.
Human Sensing as a Service
People are the sensors. A phone in the field becomes the instrument. Mobile, offline-first responses from communities turn into evidence for impact claims — honest, conservative, auditable.
Built for projects where reporting needs community input and external auditing is impractical — as a direct or auxiliary method.
Sentimus is the interface and the sensor infrastructure for community-driven verification — the data goes wherever you need it.
The rationale
Why humans, not just data?
A physical sensor measures one thing precisely until it breaks. A human sensor measures meaning, context, and change — wherever people are.
Sentiment is unreachable by sensors. Remote perception drifts from ground reality. And the people affected by an intervention have first standing to assess what's affecting them.
Where Sentimus fits
Community-side verification, without enterprise budgets.
- When reporting needs community assessment.
- Some verification questions can only be answered by the people affected. Sentimus surfaces that signal — alongside whatever hard data already exists — rather than asking funders to take the org's word for it.
- When external auditing isn't practical.
- Smaller and mid-size programs, tight budgets, limited timelines. Sentimus is meant for the situations where a six-figure third-party evaluation isn't on the table — and the claim still has to be honestly assessed.
- As a direct or auxiliary method.
- Use Sentimus as the primary verification layer, or as a complement to existing M&E. Sentimus doesn't replace what already works; it adds the community-side dimension where it's missing.
The case
Why Sentimus
Onboard once, measure for years
Once your respondents are in the network, you don't redeploy a team for every checkpoint. Re-run the pulse months later (behavior), a year out (outcome), or any replay a funder asks for — same community, fresh responses, no new fieldwork.
Phones, not platforms
Your volunteers and field coordinators don't onboard an enterprise tool. Sentimus runs on whatever phones they already carry — offline, in their language, no training cycle, no new hardware.
Your data, your stack
Responses flow out through an API into the reporting tools you already use. No platform custody, no dashboard you must log into, no exit cost if you decide to leave.
Inspect what you adopt
The methodology is in writing — formulas, weights, scoring rules, prior research cited. Audit before adopting; replace any piece (scoring, app, contract) when your context demands it.
Plugs in everywhere
REST API, webhooks, an MCP server, adapters and data flows. Scores propagate to your CRM, data warehouse, funder dashboard, or AI tooling — pullable from wherever they need to live.
The tool
The human is the sensor. The phone is the transmission.
The work of sensing belongs to people. The phone just carries the signal — a web app installed on whatever device your respondent already owns. No app-store gatekeeping, no new hardware, no enterprise platform to learn. Designed for one-handed use in sunlight, in transit, in places with patchy connection.
Audiences
Who it's for
Sentimus ships tools at the technical layer. The people below operate the process — embedding it into their own program, their own constraints, their own decisions.
Project & program managers
Running an intervention and need to verify whether it's actually working — in time to course-correct, not just to report.
Community managers & coordinators
Running the field side. Sentimus turns volunteers and local staff into a working sensor network without making them learn an enterprise platform.
Impact reporters & M&E leads
Writing for funders, boards, regulators. Sentimus produces conservative, auditable scores you can publish without hedging in the footnotes.
Technical evaluators
Reviewing the methodology and architecture before adopting Sentimus into a reporting or M&E flow. Public methodology, no proprietary lock-in, no procurement cycle needed to inspect the work.
End to end
The journey
- 1
Claim & pulse
The org publishes what's being tested and how — public, versioned, tied to funding.
- 2
Community responds
Phones in the field — anonymous, offline-capable, with photos and GPS attached.
- 3
Evidence combines
Hard records, sentiment, on-device machine signals, proximity-weighted.
- 4
Bounded score
Conservative math produces one confidence number — every input inspectable.
- 5
Activity → Behavior → Outcome
Each level a fresh pulse, months apart. Any level replayable on demand.
Refine. Replay. Repeat.
What's next
Try it in your next program — see for yourself.
We're in active build. Tell us about the program you'd run it in and we'll match it to the release wave. Or get a heads-up when the platform opens.
Optional reward layer
When orgs opt in, communities are paid for the sensing work — on Celo, in stablecoin, directly.
Sentimus ships with a reward mechanism orgs can turn on per claim. Not required — volunteer-only deployments work — but built in, ready when you want it. Paid sensing widens who responds: volunteer-only silently encodes privilege, since only people with the time and ideological alignment will participate.
- Why Celo.
- A mobile-first blockchain built for phones — phone-number wallets, stable-asset payouts (cUSD, USDC), gas low enough to pay a respondent a few cents economically.
- How it works when active.
- A
RewardPotsmart contract holds the budget for each claim that opts in. When a response meets the quality criteria — verified evidence, sufficient proximity, completed pulse — the contract pays the respondent directly. No platform custody, no settlement delay.