Training-Free Time-Series Intelligence.
Instantly.

Made for: _

Thingbook is the platform that powers DriftMind, an autonomous self-adaptive forecasting, pattern discovery and anomaly detection engine that learns online from the first data point. Available as SaaS or deployable on-prem and at the edge, with zero training, zero GPUs, and near-zero latency.

CPU-Only Architecture 140x Faster than Deep Learning Instant Adaptation to Drift
Try API Now Try CSV X-Ray
Read the technical whitepaper

Why "Smart" AI Fails at Scale

The Central Brain Trap

GenAI and Deep Learning models are "Central Brains." They are smart, but slow and expensive. They require massive GPUs, cloud round-trips, and weeks of training history. They are overkill for operational data.

The Reflex Solution

Industrial systems need "Reflexes," not brains. Thingbook uses Online Pattern Clustering and Markov-inspired Temporal Transition Graphs to memorize shapes and detect anomalies instantly. It adapts to concept drift in milliseconds, not months.

One Engine. Every Scale.

DriftMind is the only forecasting engine that deploys identically from managed cloud to a Raspberry Pi. Same API, same model, same results — the deployment target changes, the intelligence doesn't.

Cloud / SaaS

Managed platform. Start in minutes with zero infrastructure. Free tier included. Scale elastically as your data grows.

api.thingbook.io

On-Prem / Kubernetes

Full replica of the SaaS inside your infrastructure boundary. Multi-node, auto-scaling, air-gapped capable. Full data sovereignty.

Helm chart

Edge / Single Node

One Docker image, ~15 MB binary. Runs on any Linux box — factory floor, cell tower, water plant. Offline capable, no internet required.

thngbk/driftmind-edge

On-Device

Native binary runs directly on ARM or x86. Raspberry Pi, industrial gateways, embedded controllers. Autonomous forecasting with no external dependencies.

GraalVM native
Same REST API at every tier — swap the URL, keep your code
# Cloud
$ curl -X POST https://api.thingbook.io/forecasters \
  -d '{"forecasterName":"sensor","features":["temp"],"inputSize":15,"outputSize":1}'
# Edge — identical API, local execution
$ docker run -p 8080:8080 thngbk/driftmind-edge
$ curl -X POST http://localhost:8080/forecasters \
  -d '{"forecasterName":"sensor","features":["temp"],"inputSize":15,"outputSize":1}'
{"forecasterId": "e3a1...","forecasterName": "sensor"}

# On-device — same binary, no Docker needed
pi@raspberrypi:~$ ./driftmind-edge &
pi@raspberrypi:~$ curl http://localhost:8080/forecasters/e3a1.../predictions
{"anomalyScore": 0.02, "numberOfClusters": 4, "features": {...}}
Get it on Docker Hub

thngbk/driftmind-edge (API only)  |  thngbk/driftmind-edge-lab (API + Jupyter notebook)

Agent-Ready by Design

DriftMind is the first forecasting engine natively accessible to AI agents. Expose real-time predictions and anomaly scores as tools that any agent can discover and call — no integration code required.

MCP

Model Context Protocol. Claude, Cursor, Windsurf, and any MCP-compatible agent can create forecasters, push observations, and read predictions directly.

Anthropic standard

A2A

Agent-to-Agent protocol. DriftMind publishes an Agent Card so other agents discover its capabilities automatically and delegate forecasting tasks.

Google standard

REST / OpenAPI

The same API that powers SaaS, edge, and on-device. Agents use the same endpoints humans do. Full Swagger spec available for auto-discovery.

OpenAPI 3.0

From autonomous factory monitoring to agentic network assurance —
DriftMind becomes a tool in any AI agent's toolkit.

Built for Your Industry

Same engine, same API — applied to the operational challenges that define each vertical.

Telecom Networks

Real-time behavioral intelligence for network assurance, capacity planning, and service quality — integrated with your existing OSS stack.

  • Anomaly detection across RAN, core, and transport KPIs
  • Predictive capacity forecasting per cell
  • FM integration via TMF642 / TMF656
  • Deploys in under 2 weeks alongside ProOptima, InfoVista, TEOCO
Explore Telecom →

Industrial IoT

Edge-deployed predictive monitoring for manufacturing, water treatment, energy, and process industries — connected to your SCADA via standard protocols.

  • OPC-UA, MQTT, Modbus, Profinet, Profibus, MSMQ
  • Membrane fouling prediction, energy optimization
  • Predictive maintenance from vibration and pressure trends
  • PoC validated: 3–6% energy reduction in RO desalination
Explore Industrial IoT →

Data Centers

PUE optimization, thermal anomaly detection, and cooling predictive maintenance — edge-deployed alongside your existing BMS and DCIM.

  • SNMP, IPMI/Redfish, Modbus, BACnet, MQTT
  • Real-time PUE optimization setpoints
  • Thermal hotspot prediction across racks and rows
  • Cooling predictive maintenance from vibration trends
Explore Data Centers →

Built for Engineers

Dev Guide

Integration guides and performance tuning for high-throughput environments.

Swagger API

Test the cloud endpoints directly from your browser before deployment.

Python SDK

Integrate into your pipeline with `pip install driftmind-client`.

Architecture Deck

Deep dive into the Federated Edge design, TTG patterns, and benchmarks.

Work with us!

Join the team building the future of Edge intelligence. Upload your CV and tell us why you're a fit.