ProPR
Free softwareNo token markupUsage tied to tasks
ProPR and the cost of AI coding tools: every bill unbundled
You rarely pay the headline price alone: seats, model usage, cloud tasks, review volume, and the time to make agent output reviewable add up. ProPR is free under Apache 2.0 and never marks up tokens.
vs
AI coding bills
ProPR cost model
Free, open-source software on your infrastructure. Model spend goes straight to the provider through the subscriptions or API keys you already hold.
Bundled billSeats, credits, cloud tasks, or subscriptions → manual PR handoffMost tools bundle several billing shapes into one line item.
ProPR pathPlan → isolated workspace → GitHub PR → follow-up commitFree software, no token markup. Provider usage is billed to you directly.
Prices change quickly; billing shapes change slowly. A cheap editor seat gets expensive when background agents burn extra credits. A hosted agent looks simple until heavy tasks hit rate limits. A free open-source agent gets expensive when you wire every repository, sandbox, secret, retry, and PR update yourself.
ProPR is the free, self-hosted layer: plans, isolated runs, pull requests, follow-up commits, and logs. Provider spend goes to the provider. ProPR Connect is an optional hosted webhook relay for GitHub events, free for three seats and paid beyond.
Start with what the bill is buying.
Four shapes cover most of the market. Seats price people. Credits price hosted runs in the vendor's own units. API tokens price raw model usage. Subscriptions buy flat-rate access to one provider's agent. Most tools bundle several; ProPR unbundles them.
Hosted agents and AI IDEs
Usually seats, credits, premium requests, background-agent usage, cloud tasks, or provider-specific limits.
Fast adoptionProvider workspaceLimits vary
Direct agents and APIs
Subscriptions or token billing, plus the engineering time needed to turn private sessions into branches and PRs.
FlexibleManual handoffToken-sensitive
A cheap line item hides an expensive path.
| Cost source | Usually hidden in | ProPR |
|---|---|---|
| Agent/model spend | Credits, premium requests, token bills, or subscriptions | Chosen per task |
| Cloud execution | Hosted agent runs and background tasks | Your infrastructure |
| PR handoff | Manual cleanup, PR writing, and review follow-up | Built in |
| Open-source maintenance | Queues, sandboxes, logs, secrets, and retries | Shipped as the product |
Packaging agent output into a PR is the line item nobody prints.
Private sessions need packaging
Direct agents and IDEs produce useful code. Someone still cleans the branch, writes the PR, explains the change, and handles review comments.
Review-only tools stop after the comment
They cut review effort, then hand the fix back to a developer or another agent session. ProPR turns PR feedback into commits on the same branch.
Open-source stacks move cost into maintenance
A free license leaves you owning queues, sandboxes, secrets, logs, retries, PR automation, and support.
Compare the billing shape, then the number.
Seats, credits, API usage, cloud runs, PR volume, infrastructure, and the time saved when follow-up fixes land on the same PR. ProPR's own numbers live on the pricing page; the subscriptions it runs on are listed below.
SeatsCreditsPremium requestsAPI tokensCloud tasksReview volumeSelf-hosting
The subscriptions ProPR runs on, at list price.
ProPR bills nothing. Each agent draws on the plan or API key you already hold; the table shows the providers' list prices.
| Subscription | Agent in ProPR | Individual plans | Team plans |
|---|---|---|---|
| Claude | Claude Code | Pro $20/mo to Max $100/mo and Ultra $200/mo | Team $25–$125/seat/mo |
| ChatGPT | Codex | Plus $20/mo to Pro $100–$500/mo | Business $25/user/mo |
| Google AI | Antigravity CLI | Free tier; Pro $19.99/mo to Ultra $99.99–$200/mo | Google Cloud plans from $30/seat/mo |
| Mistral Vibe | Mistral Vibe | Free tier; Pro $14.99/mo | Team $24.99/user/mo |
| OpenCode | OpenCode | Free models; Go $10–$40/mo, or pay per token | Go Teams $20/member/mo |
Monthly billing, USD, before tax; annual billing is cheaper on most plans. Every provider also sells API usage pay-as-you-go, metered per token outside these prices. Verified November 2026: Claude pricing, Codex pricing, Antigravity pricing, Google AI subscriptions, Mistral pricing, OpenCode Go.
Low sticker prices hide high total costs.
Agentic coding burns money beyond seats: long contexts, repeated review cycles, branch cleanup, manual PR writing, failed attempts, and cloud task limits.
- Ask how a long-running task is billed.
- Ask whether retries and follow-up changes consume the same quota.
- Ask who pays for compute, storage, and logs.
ProPR makes the cost boundary explicit.
You choose the agent and model per phase and switch whenever pricing or model quality shifts, with every call in the usage log.
- Use subscription-backed agents where they cost less.
- Use API billing where direct usage control matters.
- Switch models without changing the PR path.