Welcome to your workspace!
Connect your code repo, run a scan, and search for the right cloud GPUs.
Workspace
Sign in, add repos, and prep scans.
Repositories
2 reposPublic GitHub repositories only.
Scan controls
Run a scan to automatically generate a workload profile.
Select a repository to scan.
Workload profile
Tell us how you plan to run this job so we can size GPUs and rank marketplace results.
You can adjust GPU + VRAM filters after the scan in Marketplace search.
Results
Live scan output and recommendations.
Codebase scan
IdleDetected workload
Needs reviewJob type
—
Model size
—
Dataset size
—
Run length
—
GPU count
—
VRAM per GPU
—
Total VRAM
—
Max $/hr
—
Total budget
—
Cost preference
—
Risk tolerance
—
Spot allowed
—
Does this look right?
Detected entrypoint
Needs reviewSelect the command that actually launches your job.
Choose one to lock recommendations.
Review extracted parameters
OptionalConfirm or edit key training inputs. These override inferred values.
Overrides are stored locally for this repo.
AI suggestions
OptionalAI suggests missing workload fields only. Apply only if they look correct.
No AI suggestions available.
Audit trail
Why the scanner inferred these values.
No evidence available.
Risk analysis
AutoScan JSON
Demo repo code
main.pyMarketplace search
Search using scan output and deep filters.
Providers currently supported: Vast.ai and Runpod.
Search filters
CustomGPU + pricing
System requirements
Runpod does not expose disk size; disk filtering is disabled.
Ranking
Detected workload hints can be pulled into search filters anytime.
Market results
LiveRun a search to see marketplace inventory.