Investor brief / September 2026 — reciprocal network model
Senlay tells humans and machines what is happening at a specific place, right now — and shows how much that answer can be trusted.
Coordinate-level environmental intelligence with uncertainty at its centre. Contribute a sensor where no network reached and you get the whole layer in return. Consume without contributing and you pay.
This is not a pivot away from the API. It is a change in how the API is fed and who pays for it. The verification engine, the trust scoring and the endpoints are unchanged and already in production.
Weather, marine, aviation, air quality, terrain, satellite, hydrology, and hazard sources — the raw material a network reasons over, not the moat itself.
Protocol, JS SDK, MQTT bridge, and edge gateway published Apache 2.0. Provider/station registration and HMAC-signed ingestion validate in production today.
ok, no_data, out_of_coverage, not_configured, error, timeout. A source dark for a missing credential is labelled as such, never disguised as inapplicable.
Said plainly. One person, including Linux deployment. The gap this brief exists to close is external usage from people I have never met.
The problem
Operational agents need evidence, not just forecast values.
Agents are starting to plan drone missions, schedule outdoor crews, route vessels, time spraying, and support real-world decisions. Generic forecasts and LLM guesses are not enough when wind, waves, weather, terrain, air quality, and sensor freshness can change the answer.
Training data cannot answer "right now."
An LLM can sound confident while using stale, generic, or cached context for a local decision.
Generic weather APIs were built for broad dashboards.
They return values, but usually do not expose source hierarchy, freshness, local disagreement, or domain meaning.
The missing layer is local verification for outdoor decisions.
Agents need measurable conditions with source, distance, confidence, and a plain-language interpretation they can use in a workflow.
Product
One verification request returns conditions, evidence, risk context, and decision meaning.
Senlay is not a consumer weather app. It is infrastructure for SmartSurf, outdoor AI apps, IoT systems, and safety applications that need source-aware physical verification before they act.
Measured conditions
Wind, gusts, weather, waves, tides, currents, air quality, terrain, hazards, and sensor availability; bathymetry only when a validated layer is configured.
Evidence per value
Source type, station distance, measured timestamp, freshness, trust score, and whether the value is sensor or model derived.
Evidence-ready interpretation
Structured verification output for safety workflows, application code, and audit trails.
Named, honestly — a reviewer will find these in five minutes
Senlay is not another sensor network. It is the verification layer above all of them.
Do not lead with network density — that race is already lost on density and does not need to be run. Contributed sensor networks already exist and some are large. They are not competitors. They are inputs.
Open-Meteo, Tomorrow.io, Stormglass, Visual Crossing.
Penalized for admitting missing coverage, so their APIs interpolate across gaps and present the result as a measurement. In safety-critical use that is worse than no answer.
Synoptic Data — the closest technical neighbour.
Aggregates hundreds of networks without owning stations and flags physically plausible outliers. Synoptic cleans data for analysts. Senlay reasons for systems that act.
WeatherXM — roughly 10,000 community stations, 81 countries, ~$13M raised.
A good weather network, rewarding owners in tokens for hardware it manufactures. Senlay is not trying to out-deploy it. WeatherXM is a source family, one input among many reasoned over.
Archetype AI — $13M seed, $35M Series A (Nov 2025, IAG Capital / Hitachi Ventures, Bezos Expeditions, Venrock, Amazon, Samsung).
Newton TimeFusion reasons over sensor streams for an enterprise's own assets inside its own perimeter. Senlay is public outdoor context at a coordinate, aggregated across networks nobody owns, aimed at safety decisions nobody has a commercial reason to open up otherwise.
Senlay's position is one layer up: the place where readings from every network are cross-checked against each other and against forecast models, weighted by distance, freshness and calibration, and returned with the disagreement made explicit rather than averaged away.
Evidence Object v1
A proof block attached to physical-world claims.
Evidence Object v1 is the funded milestone that makes each important Senlay value auditable: where it came from, when it was measured, how far away it is, whether sources agree, and why the interpretation was made.
"Wind is risky."
The agent receives a conclusion but cannot inspect whether the source was live, nearby, fresh, or trusted.
"Wind is risky, and here is why."
The answer carries source, timestamp, distance, measured/model status, confidence, disagreement, and interpretation reason.
{
"wind.speed_ms": {
"value": 9.4,
"unit": "m/s",
"evidence": {
"source_type": "sensor",
"age_min": 6,
"distance_km": 4.2,
"confidence": "medium-high"
},
"disagreement": {
"alternate_source": "model",
"delta": 3.3,
"weighting": "live_preferred"
}
}
}Market and wedge
Water sports is the wedge, not the ceiling.
SmartSurf is the first proving ground: the domain where I can tell instantly whether the answer is true, after twenty years reading wind and water for other people's safety. The same evidence engine reasons for any system that acts outside and would never install a weather station itself.
SmartSurf schools and stations
Clear physical risk, real hardware path, obvious buyer pain, and a visible workflow from GPS trigger to Senlay context to station alert.
AI agents for drones, field ops, agriculture
Weather, heat, wind, humidity, air quality, and terrain affect launch, scheduling, spraying, work safety, and logistics.
Marine, SAR, insurance, autonomy
Higher-value workflows after reliability, evidence contracts, SLAs, and domain validation mature.
Business model
Contribute or pay. The two sides do not compete for the same budget.
Attach a sensor, receive the network.
No cash. They are coverage and distribution, not revenue.
Usage-based, self-serve.
Outdoor apps, IoT products, agent frameworks. Free tier with a daily allowance.
Structured risk events, not raw context.
Schools, stations, fleet operators. Higher value per seat — the layer SmartSurf validates.
Parametric insurance, precision agriculture, drone ops.
Verified evidence with an audit trail is a compliance requirement and priced accordingly.
Reciprocal fleet integration.
Sensor-network operators and hardware vendors integrate fleets, gaining verification and reach.
Bring your own sources and your own model.
Their data contracts stay in place; Senlay supplies the trust layer above them. Direction, not a committed timeline.
Modes 1–4 are the current product direction: the live beta supports self-serve exploration, while paid pilots and enterprise delivery remain founder-led. Modes 5 and 6 are longer-term enterprise motions with different sales cycles — presented as direction, with no timeline attached. Revenue comes from systems that act and will never install hardware; coverage comes from people who already own instruments and want better answers about their own location.
Why this compounds
Contributed data quality is a risk we make visible and test explicitly.
Instruments can end up beside air-conditioning units, in direct sun, or uncalibrated. Senlay is designed to compare source distance, freshness, and model/hardware disagreement and to expose gaps rather than silently present a consensus as ground truth. The contributor calibration and neighbouring-station checks are being validated through pilots; a poor sensor must be treated as evidence with uncertainty, not assumed reliable.
Documentation moves the starting point. Behaviour moves the weight, permanently.
A certified instrument fails like any other. An undocumented one that tracks its neighbours for months gains weight without any paperwork. This is deliberately not purchasable — the answer to spoofing.
What compounds is not station count. It is place memory.
Eligible requests can store a compact snapshot for a coordinate. The current system returns descriptive historical comparisons when enough paired observations exist; whether that improves decisions is being measured in held-out evaluations, not assumed from request count alone.
Data sources and language models are interchangeable. The judgement is not.
Normalization, distance/freshness scoring, measured-vs-predicted separation, weighting, divergence detection, and refusal to infer all live in one place: the trust engine. It stays closed and non-configurable by design.
Giving network access in exchange for data attacks a data vendor's own pricing.
For Senlay it is the growth mechanism. Open protocol, closed trust engine — shipped, not planned.
Traction and roadmap
Live platform, public beta, and a path to paid pilots.
| Stage | Milestones |
|---|---|
| Live today | Public site, API docs, interactive demo, developer dashboard, BYOK LLM profile keys, SmartSurf integration path, risk_event endpoints, and Physical World Model responses with source/freshness/confidence concepts. |
| Next 6 months | Evidence Object v1, drone go/no-go domain pack, outdoor-work domain pack, reliability and monitoring, data quality scoring, and 5-10 paid pilots. |
| Next 12 months | 25-50 paying developer/team customers, 3-5 enterprise pilot conversations or contracts, agriculture domain pack, and private sensor integration path validated. |
19-slide rebuild
Old pitch logic plus new live-evidence story.
SENLAY / investor deck / rebuilt in current style
Slide deck
The full deck now combines both source sets.
The first 12 slides rebuild the reviewed pitch deck. The last 7 slides rebuild the newer investor-brief PDF around live evidence, hardware/model divergence, sensor hierarchy, and founder velocity.
- 01-12
Problem, insight, product, demo, why now, wedge, pricing, moat, founder, roadmap, ask.
- 13-19
Live evidence update, Hoi An case, forecast trust, sensor hierarchy, agent-verified fixes, thesis.
Founder-market fit
20 years reading wind, water, and risk in real conditions.
Viktor Kryvotsiuk is the sole human founder in Hoi An, Vietnam. Senlay grew out of water-sports instruction, marine safety decisions, and SmartSurf hardware work before becoming a horizontal API for agents.
Viktor Kryvotsiuk
- 20+ years certified kitesurfing and water-sports instruction across Ukraine, Vietnam, Egypt, and Sri Lanka.
- IYT International Bareboat Skipper, Power & Sail up to 24 m.
- 1NCE Certified IoT Integrator.
- Built SmartSurf GPS tracker hardware for riders and boards.
- Shipped Senlay as a live public beta with an AI-agent coding team.
Investor FAQ
Hard questions, direct answers.
Is this just another weather API?
No. Weather APIs return values for dashboards. Senlay returns evidence-bearing context for agent reasoning: source, freshness, distance, confidence, disagreement, and domain interpretation.
If contributors don't pay, who does?
Systems that actuate commercial decisions and will never contribute hardware: drone and BVLOS operators, robotics fleets, lifting and construction, parametric insurers, marine operations. They need the answer, not the network.
Water-sports schools and stations validate the operational-safety tier through SmartSurf first — the proving ground, not the ceiling.
Isn't this just a failed network four years late? WeatherXM already has 10,000 stations.
The most dangerous misreading. WeatherXM is an input, not an opponent — Senlay reasons across networks rather than owning one. The open contributor protocol exists for the places no network reached; density is raw material, not the target. Never lead with node counts.
Why not sell to every vertical immediately?
The first wedge is narrow on purpose. Drone go/no-go and field operations are concrete enough to demo quickly. The horizontal market comes from adding domain packs on the same evidence layer.
Could a large platform build this?
They could build pieces. The hard work is fragmented, operational, and domain-specific: source normalization, freshness, confidence, live/model disagreement, domain rules, and field interpretation.
How does Senlay handle safety?
Senlay is not a safety authority. It provides auditable context. The customer application or operator makes the final decision, which is why source, freshness, confidence, and limits matter.
$500K pre-seed to prove the reciprocal network with real external usage.
Use of funds: signed evidence receipts, contributor onboarding and the OGC SensorThings adapter, API reliability and monitoring, SmartSurf risk-event proof as the operational-safety tier, domain packs, and senior backend support. Pre-revenue, no paying customers — said plainly.