Four fictional monthly tools over 12 months: (20 + 12 + 15 + 10) x 12.
Which subscriptions actually earn a place?
Use your own costs and capability labels. This tool does not fetch prices or judge model quality; it solves the portfolio you describe.
Subscription receipt tape
rows 0errors 0
Current horizon cost$0
Exact optimized cost$0
Computed savings$0
Cost compression
0 DP states
awaiting data
awaiting data
Required jobs
| Decision | Subscription | Billing | Price | Horizon cost | Capabilities | Coverage |
|---|
A bundle wins only when coverage survives.
Count is not cost. Cost is not coverage. Normalize every billing cadence, define the jobs that matter, then minimize spend while preserving the complete required capability set.
dp[newMask] = min(dp[newMask], dp[mask] + horizonCost(subscription))
One fictional lifetime candidate covering all five required jobs within the horizon.
Exact computed difference, not a claimed live deal or quality judgment.
Know what the solver knows, and what it cannot.
Billing normalization
Monthly price multiplies by horizon months. Annual price multiplies by each started 12-month period. Lifetime price is charged once inside the horizon.
Coverage truth
The solver trusts your pipe-delimited capability labels. A bundle missing research is incomplete when research is required, regardless of its sticker price.
Decision limits
Taxes, usage caps, quality, privacy, switching cost, and changing prices are not inferred. Add those considerations to your source data before acting.