ESPectre

Roadmap

Release unavailable

Follow ESPectre from one node sensing platform to multi-node orchestration.

Roadmap · Project direction

From a single sensing node to a coordinated local platform.

ESPectre is evolving toward a shared platform for broader integrations, new sensing capabilities, and coordination between nearby nodes. Local operation remains the baseline, while new capabilities enter the product only after hardware support and sensing evidence are in place.

Local by default Motion, not media Built to embed
Planned · September 2026 v3.0.0
A shared foundation for product integration

Make the sensing layer dependable and easy to carry into products.

Version 3.0.0 is planned as the shared baseline for ESPectre as a reusable platform. One shared sensing core supports several firmware paths, so product teams can add local motion sensing without maintaining a separate implementation for each integration.

Shared architecture
The sensing core and shared runtime intended for the stable v3 baseline.
Runtime contracts
Runtime behavior and protocol services frozen as one platform contract.
Supported frontends
ESPHome, Native, and Matter included in the candidate scope.
Release surfaces
Release artifacts and the embeddable SDK included in whole-platform validation.

Put private sensing inside real products.

The v3 line brings ESPectre closer to real products with broader integrations, SDK support, and new sensing capabilities.

  1. v3.1.0Expand Matter support and validate it across more controllers.
  2. v3.2.0Bring ESPectre to Arduino projects through a supported SDK runtime.
  3. v3.3.0Bring ESPectre into Apple Home with a dedicated HomeKit frontend.
  4. v3.4.0Recognize stationary presence, even when no movement is visible.
  5. v3.5.0Explore gesture control and breathing-related micro-motion.
Planned · 2027 v4.0.0

Accelerate sensing and let nearby nodes work as one system.

The planned v4 platform combines local node coordination, hardware-accelerated sensing, and an optional, self-hostable WebSocket relay. Devices connect outward through authenticated WSS, browsers use WSS to the same service, and local Direct HTTP remains the independent default.

  • Discover peers, coordinate traffic, and degrade cleanly when a node disappears.
  • Use hardware acceleration to process more CSI, run more detectors at once, and preserve headroom for networking and coordination.
  • Claim and manage rooms and devices, view live and historical state, and configure alerts.
  • Choose local, self-hosted, or managed relay.espectre.dev deployment without exporting raw CSI.
  • Gate relay launch on pairing, per-device credentials, revocation, tenant isolation, bounded queues, heartbeat, reconnect, rate limits, and a reviewed threat model.
Subject to hardware availability v5.0.0

Build on Wi-Fi hardware designed for sensing.

v5 starts when practical embedded hardware exposes IEEE 802.11bf or equivalent measurements. Deliberate sensing procedures could make radio data more repeatable, improve coordination between nodes, and open a path to richer presence and motion capabilities.

  • Qualify the first supported radio, driver, and measurement API on practical embedded hardware.
  • Add a standards-backed sensing backend and validate every detector against new datasets.
  • Measure calibration, false positives, stationary presence, and multi-node fusion from scratch.
  • Let existing products adopt the new backend through the same SDK and protocol, with a documented migration path from ESP32 CSI.
RESEARCH TRACKS

Research areas under evaluation.

Research does not reserve release scope. The tracks are listed in dependency order, and each may become a product, remain a host-side experiment, or end with a measured rejection.

01

HE20 and HT40 sensing profiles

Map both layouts onto canonical detector inputs, compare their benefits and costs, and select any profile that earns promotion.

02

Higher CSI rate

Select the highest useful sustained rate within declared limits for loss, jitter, compute, memory, and transport load.

03

Longer and multi-scale windows

Test slow micro-motion over longer windows while preserving the current movement response.

04

Stationary presence

Use the selected capture profile to test whether an occupied quiet room can be distinguished from an empty one.

05

Breathing-related motion

Explore non-medical breathing-related micro-motion after stationary presence and longer windows are validated.

06

Brief gestures

Build a gesture-specific corpus after the higher-rate path preserves the necessary short-timescale information.