Pricing¶
SC-NeuroCore is dual-licensed: open-source for research and education, commercial licenses for proprietary integration.
Early Adopter Pricing — First 25 Customers
SC-NeuroCore commercial licenses are available at introductory pricing for the first 25 customers. These rates lock in permanently — when the program closes, standard pricing applies to new customers.
Spots remaining: 25 of 25
Community
- 122 neuron models (113 biophysical + 9 AI-optimized)
- Rust SIMD engine (41.3 Gbit/s AVX-512)
- 19 Verilog HDL modules + 67 formal properties
- ODE → Verilog compiler (
sc-neurocore compile) - PyTorch training (7 surrogates, 10 cells)
- 6-codec neural data compression
- 125-function spike analysis toolkit
- Quantum hybrid (Qiskit/PennyLane)
- 84 tutorials + full API documentation
- Community support (GitHub Discussions)
- Source modifications must remain open (AGPL)
Professional
First 25 customers- Everything in Community
- Closed-source integration permitted
- Priority email support (48h business hours)
- FPGA synthesis support + build guidance
- Visual SNN Design Studio (coming soon)
- Custom neuron model development (2 models/year)
- Quarterly security advisories
- No AGPL copyleft obligation
Enterprise
First 25 customers- Everything in Professional
- Unlimited seats across organization
- Priority email support (24h business hours)
- On-premise deployment assistance
- Custom FPGA target integration + RTL adaptation
- Formal verification reports for your design
- White-label and OEM licensing
- Joint development agreements
Founding Member
- Everything in Professional
- Lifetime price lock (CHF 290/yr — even when standard is CHF 1,490)
- Direct access to lead developer (email + video calls)
- Input on roadmap priorities
- Name in CONTRIBUTORS.md + release notes
- Free 30-day evaluation before commitment
- Early access to Visual SNN Design Studio
Spots remaining: 10 of 10. When gone, this tier closes permanently.
What You Get¶
ODE → FPGA in One Command¶
sc-neurocore compile "dv/dt = -(v-E_L)/tau + I/C" \
--threshold "v > -50" --reset "v = -65" \
--params "E_L=-65,tau_m=10,C=1" --init "v=-65" \
--target ice40 --testbench --synthesize
Write neuron equations as strings. The compiler produces synthesizable Q8.8 Verilog RTL with saturating arithmetic, transcendental function LUTs (exp, log, tanh, sigmoid, sin, cos), auto-generated testbenches, and one-click Yosys synthesis. Python simulation matches Verilog bit-for-bit (67 formal properties verified by SymbiYosys).
PyTorch Training → SC Bitstream → FPGA¶
7 surrogate gradient functions, 10 differentiable neuron cells
(nn.Module), SpikingNet and ConvSpikingNet architectures. Train on
GPU with standard PyTorch optimizers, export weights to stochastic
computing bitstreams via to_sc_weights(), compile to Verilog RTL.
Trainable per-synapse delays (DelayLinear).
6-Codec Neural Data Compression¶
ISI+Huffman, Predictive (4 learnable predictors), Delta, Streaming, AER,
and WaveformCodec (24x compression on 1024-channel Neuralink-scale data).
Unified API: get_codec(name), recommend_codec(). Rust backend (780x).
122 Neuron Models + Rust SIMD Engine¶
82 years of computational neuroscience (1943-2026). 111 Rust models with PyO3 bindings, 81-model NetworkRunner with Rayon-parallel populations. 41.3 Gbit/s bitstream packing (AVX-512). 224 Mstep/s LIF throughput.
125-Function Spike Analysis Toolkit¶
CV, Fano factor, cross-correlation, Victor-Purpura distance, SPIKE-sync, Granger causality, GPFA, SPADE pattern detection. Matches Elephant + PySpike combined. Pure NumPy — no additional dependencies.
Feature Comparison¶
| Feature | Community | Professional | Enterprise |
|---|---|---|---|
| 122 neuron models + Rust engine | Yes | Yes | Yes |
sc-neurocore compile (ODE → Verilog) |
Yes | Yes | Yes |
| PyTorch training (7 surrogates, 10 cells) | Yes | Yes | Yes |
| 6-codec neural compression | Yes | Yes | Yes |
| 125-function spike analysis | Yes | Yes | Yes |
| 19 Verilog modules + 67 formal properties | Yes | Yes | Yes |
| Quantum hybrid (Qiskit/PennyLane) | Yes | Yes | Yes |
| 84 tutorials + full docs | Yes | Yes | Yes |
| Closed-source use | No (AGPL) | Yes | Yes |
| Priority support | Community | 48h | 24h |
| FPGA synthesis support | Self-serve | Guided | Custom targets |
| Visual SNN Design Studio | — | Yes (coming) | Yes (coming) |
| Custom neuron models | — | 2/year | Unlimited |
| Formal verification reports | — | — | Yes |
| OEM / white-label | — | — | Yes |
| On-premise deployment | — | — | Yes |
Why Buy Now?¶
Standard pricing (after early adopter program closes):
Professional: CHF 1,490 /yr per seat
Enterprise: CHF 14,900 /yr site license
Early adopter pricing (first 25 customers — locked permanently):
Professional: CHF 490 /yr per seat ← you save CHF 1,000/yr
Enterprise: CHF 4,900 /yr site license ← you save CHF 10,000/yr
Founding Member: CHF 290 /yr per seat ← you save CHF 1,200/yr
Example: Year 1 you pay CHF 490.
Year 5 new customers pay CHF 1,490.
You still pay CHF 490. Every year. Forever.
SC-NeuroCore is the only open-source framework with a complete ODE → PyTorch training → SC bitstream → Verilog RTL → FPGA bitstream pipeline. The standard pricing reflects this — CHF 1,490/yr is still less than a single Vivado license (CHF 3,000+/yr), and SC-NeuroCore includes the full SNN training stack, 122 neuron models, and formal verification that Vivado doesn't offer.
Early adopters get this at 67-81% off. Permanently.
Academic Pricing¶
Free Professional license for .edu email addresses. Includes closed-source rights for thesis work and research prototypes. No strings attached.
FAQ¶
Can I use the Community edition for commercial research? Yes, as long as your modifications are also released under AGPL-3.0. If you need to keep your code proprietary, choose Professional.
What FPGA targets are supported?
The RTL is vendor-agnostic (Verilog-2005). The compile CLI targets
ice40, ECP5, Artix-7, and Zynq. Professional includes guided synthesis
support. Enterprise adds custom ASIC targets.
What is the early adopter program? The first 25 paying customers lock in current pricing permanently. When the 25th customer signs up, prices increase to standard rates for all new customers. Existing customers keep their locked rate.
What is the Founding Member tier? 10 spots at 81% off standard pricing — CHF 290/yr instead of CHF 1,490. Includes direct developer access, roadmap input, and a free 30-day evaluation. Once 10 spots fill, the tier closes permanently.
Can I evaluate before purchasing? The Community edition is fully functional. Founding Members additionally get 30 days free before their first payment.
What is the Visual SNN Design Studio? A web-based IDE for visual SNN design — equation editor, network canvas, training monitor, compiler inspector, synthesis dashboard. Currently in development. Professional and Enterprise licensees get access at launch.
What payment methods do you accept? Credit card via Polar.sh, bank transfer (IBAN), or invoice (NET-30 for Enterprise). All prices in CHF. EUR and GBP accepted at daily exchange rate.
Do you offer refunds? 30-day money-back guarantee on all tiers. No questions asked.
Ready to deploy neuromorphic AI on silicon?
Get Started — CHF 490/yr
Talk to Us
SC-NeuroCore is developed by ANULUM — advancing neuromorphic computing from simulation to silicon.
Contact: protoscience@anulum.li | www.anulum.li | ORCID: 0009-0009-3560-0851