Pricing
Codality is $89/year per instance. One license, unlimited usage. No tiers, no per-seat charges.
During the beta period, no license key is required.
What the license includes
- One server instance — Run Codality on one server
- Unlimited projects, tickets, and agent runs
- Unlimited users
- All features — No tiers, no feature gates
- All updates released during the license period
What you pay separately
| Cost | Provider | Notes |
|---|---|---|
| OpenAI API usage | OpenAI | You bring your own API key. Token costs depend on model and usage. |
| SendGrid | SendGrid | For OTP login emails. The free tier (100 emails/day) is sufficient for most teams. |
| Tavily | Tavily | For web search and content extraction tools. Free tier available at tavily.com. |
| Server hosting | Your choice | Any Linux server. A $5–$20/month VPS is sufficient. |
Estimated API costs per model
Codality tracks token usage per agent session. Costs depend on the model and session length.
| Model | Input (per 1M tokens) | Output (per 1M tokens) | Notes |
|---|---|---|---|
gpt-5.2 |
$2.00 | $8.00 | Default. Strong general-purpose model. |
gpt-5.1-codex |
$0.50 | $2.00 | Optimized for code. Good value. |
gpt-5.3-codex |
$0.50 | $2.00 | Latest codex variant. |
gpt-4.1 |
$2.00 | $8.00 | Capable, mature. |
gpt-4.1-nano |
$0.10 | $0.40 | Used for fast model tasks (plan readiness, titles). |
o3 |
$2.00 | $8.00 | Strong reasoning. |
o4-mini |
$1.10 | $4.40 | Reasoning at lower cost. |
A typical agent session uses 50K–500K input tokens and 2K–20K output tokens. At gpt-5.2 rates, that's roughly $0.10–$1.00 per session. Complex tasks with long sessions can use more.
Beta period
Codality is currently in beta. During this period, no license key is required. All features are available.
When beta ends:
- A license key will be required for new installations
- Existing installations receive a grace period to purchase
- No data is lost or locked if you don’t purchase immediately
- The license is $89/year, renewable annually
Self-hosted
Codality runs entirely on your infrastructure. Your code never leaves your server. The only external calls are:
- OpenAI API — for AI operations
- SendGrid — for OTP emails
- GitHub API — for integration (optional)
No telemetry is collected without explicit consent.