Configure
Build RAW-to-output pipelines.
Select capture, Bayer/RAW handling, preprocessing, ISP, inference handoff, fusion, and output stages as an explicit SDK pipeline.
- RAW/ISP stage graph
- Typed module configuration
- Reusable camera presets
ACCELERATED ISP PIPELINES FROM RAW PIXELS TO VISION OUTPUT
At Turing Intelligence, we design and build end to end vision systems.
Our vision: "Converting Complex Problems into Easier Computations by Intelligence"
Turing Intelligence is an engineering company based in Prague, we develop software for computer vision, sensor processing, and GPU accelerated edge systems.
- Image Signal Processing
- GPU Accelerated ISP
- Computer Vision
- Edge Computing
- Embedded Vision Systems
Product
One SDK, from sensor to deployment.
Built for RAW sensor and ISP workloads, Turing Signal ProcessingTM helps engineering teams compose, profile, and deploy latency sensitive image pipelines from capture to vision output.
Profile
See where every frame spends time.
Measure per-stage latency, throughput, memory movement, and GPU work so ISP pipelines can be tuned before deployment.
- Per-stage latency traces
- GPU throughput telemetry
- Deployment readiness evidence
Deploy
Run consistently on edge GPUs.
Keep the same SDK pipeline model across NVIDIA Jetson targets and Linux GPU workstations for controlled production rollout.
- Binary SDK delivery
- Target-specific ISP tuning
- Controlled edge rollout
Native C++
CUDA Acceleration
Per-Stage Pipeline Profiling
Jetson + Linux GPU
Pricing
Start with evidence.
Scale with confidence.
Choose the plan that matches according to your needs and requirements.
Evaluation
€0/ 30 days
Validate compatibility before committing to a paid engagement.
- Binary SDK and examples
- One named developer
- One target environment
- Non-production use
Recommended first step
Technical Pilot
€2,900fixed-scope pilot
A focused integration pilot for one sensor pipeline and target platform.
- 90-day pilot access
- Technical onboarding session
- Reference pipeline configuration
- Integration review and findings
Production
€3,900/ month
For a small engineering team and an initial controlled deployment.
Monthly subscription shown. Yearly option: €40,000 / year.
- Up to 5 developer seats
- Up to 10 deployed SDK nodes
- Updates during subscription
- Standard technical support
OEM and high-volume deployment
Custom
Custom monthly or yearly subscription terms for higher deployment volumes, custom modules, or commercial redistribution.
- Volume-based SDK terms
- Custom module development
- Support and SLA options
- Source escrow by agreement
Indicative prices exclude VAT, hardware, travel, and project-specific engineering. Final scope depends on target hardware, required modules, support level, and deployment volume. TSP SDK access is provided through subscription or pilot terms.
Questions
Before an evaluation.
What information is needed to qualify a target?
We normally need the sensor or input format, target compute platform, resolution or sample rate, required operations, latency or throughput target, output interface, and deployment volume.
Is this a cloud service?
No. The TSP SDK is intended for local execution on the target system. Cloud services can be added separately for device management, telemetry, or dataset workflows.
Does the evaluation include production rights?
No. The evaluation package is for compatibility and technical assessment. Production deployment requires a production or OEM license.
Are all modules available on every target?
Module availability and performance depend on the target hardware, input format, and required pipeline. The support matrix is confirmed during evaluation or pilot scoping.
Is source code included?
The standard packages are binary SDK licenses. Source access, source escrow, and custom ownership terms can be negotiated for OEM or custom engineering engagements.
Contact
Let’s build together.
Tell us what you’re building, the signals you’re working with, and where the system needs to run. We’ll review the constraints and respond with a practical engineering path.