Formal verification
Use the service as a fallback lane when internal compute is saturated or when you want external reproducible runs on uploaded CNF instances.
Logic Workloads
Online SAT solver for large DIMACS CNF jobs
Use Both Hemispheres when you need SAT throughput without running your own solver fleet. Upload a DIMACS CNF file, queue a single-threaded job with a one-hour limit, and only pay to view or download the satisfying assignment when one exists.
Who It Is For
Overflow capacity for serious CNF workloads
Research groups
Run large combinatorics, cryptanalysis, or benchmarking instances without maintaining your own web upload flow and payment collection.
Consultants and boutique teams
Offer solver-backed analysis to clients without building your own SAT job queue, storage, billing, and result-delivery path.
How It Works
Simple pricing and clear queue behavior
1. Upload
Send a `.cnf` file in DIMACS format through the web interface. Files up to `100 MB` are accepted.
2. Solve
Jobs run on worker machines with one solver thread and a hard time limit of `1 hour`.
3. Review result
`UNKNOWN` and `UNSATISFIABLE` results are visible for free. If a satisfying assignment exists, you can unlock it for `$10`.
Reduce Friction
Try the flow with sample CNF files and a published benchmark
Sample files
Start with tiny known-good examples before uploading a larger production instance.
Public benchmark
Review the measured March 30, 2026 SAT smoke benchmark taken against the public service before you run your own workload.
Why Buyers Choose It
Reasoning infrastructure without building the wrapper
- Direct web upload flow for technical users who already have CNF instances.
- Clear queue semantics with single-thread runs and explicit resource limits.
- Pay only for the valuable artifact: the satisfying assignment.
- Human support for solver questions and unusual workloads.
Next Step
Run a real CNF job on the live service
Start with trial credits, upload one instance, and use the support address below if you want help benchmarking a workload or discussing recurring usage.