We publish what others hide. Every number is reproducible.
Every tool compiles to a single static binary. No runtime dependencies, no shared libraries, no container overhead.
| Tool | Binary | Docker | Competitor | Competitor size |
|---|---|---|---|---|
| Bounty | 14MB | 18MB | Jira | 2.1GB (Docker) |
| Headcount | 13MB | 17MB | Plausible | 890MB (Docker) |
| Salt Lick | 12MB | 16MB | Flagsmith | 1.4GB (Docker) |
| Corral | 9MB | 13MB | Svix | 650MB (Docker) |
| Strongbox | 13MB | 17MB | Vault | 380MB (Docker) |
| Bellwether | 12MB | 16MB | Uptime Kuma | 420MB (Docker) |
Time from binary execution to first HTTP response. Measured on a $5 VPS (1 vCPU, 1GB RAM).
| Tool | Stockyard | Competitor | Competitor startup |
|---|---|---|---|
| Bounty | 38ms | Jira | 45-90 seconds |
| Headcount | 35ms | Plausible | 12-18 seconds |
| Salt Lick | 32ms | Flagsmith | 8-15 seconds |
| Strongbox | 40ms | Vault | 3-5 seconds |
RSS memory after startup with an empty database. Measured with no active connections.
| Tool | Stockyard | Competitor | Competitor memory |
|---|---|---|---|
| Bounty | 11MB | Jira | 2-4GB |
| Headcount | 12MB | Plausible | 200-400MB |
| Salt Lick | 10MB | Flagsmith | 300-500MB |
| Corral | 9MB | Svix | 150-250MB |
| Strongbox | 13MB | Vault | 50-100MB |
SQLite with WAL mode. Single file, no server process, no connection pooling, no vacuuming needed.
| Records | DB size | Write latency | Read latency |
|---|---|---|---|
| 1,000 | 180KB | 0.2ms | 0.05ms |
| 10,000 | 1.8MB | 0.3ms | 0.08ms |
| 100,000 | 18MB | 0.5ms | 0.12ms |
| 1,000,000 | 180MB | 0.8ms | 0.2ms |
Benchmarks measured on Ubuntu 24.04, 1 vCPU, 1GB RAM (Hetzner CX11). Competitor numbers from official docs and community reports. Last updated: April 2026.