Hosted Solo
€59/ month
- Developer seats
- 1
- Concurrent agent runs
- 2
- Data storage
- 20 GB
- Backup retention
- 60 days
ProPR Hosted runs the full ProPR backend in a dedicated VM that the ProPR team operates. Connect handles GitHub routing, billing, and lifecycle status. The fleet service provisions, checks, suspends, backs up, and retires the VM through operator-approved plans.
Limited capacity. Hosted is open for paid signups. Capacity is limited, and Connect checkout shows whether a new instance is available or sold out.
t-<instance>.propr.devNo public inbound ports. Management access accepts only the operator network.
Each plan includes the managed ProPR backend, Connect routing, a managed UI tunnel, and encrypted application backups. Repositories have no count limit; they share the plan's storage. Prices are monthly, in euros, and exclude VAT where applicable.
€59/ month
€149/ month
€349/ month
Additional Hosted seats cost €15 per seat per month and raise the seat count only. Concurrency, storage, and backup retention follow the plan. Connect Plus seats for self-hosted stacks keep their own USD pricing on the pricing page.
A seat is a distinct GitHub developer who starts ProPR work through Connect in a billing cycle. Bots such as GitHub Actions and the ProPR bot never take a seat. At the limit, Connect replies on the issue or pull request and starts no task for new developers. Connect applies the larger of two allowances: your Connect allowance (3 free users, or your Plus seats) or your Hosted seat total, which is the plan's included seats plus any additional Hosted seats. The allowances never add together, so Hosted Solo keeps the 3 free Connect users.
The plan sets the ProPR worker concurrency on your VM, and the setting cannot be raised from the ProPR UI. Additional tasks wait in the queue until a run finishes.
Data storage holds repository clones, agent workspaces, the database, logs, and settings. Each VM also gets a fixed 20 GB system reserve for the operating system and the ProPR runtime, which does not count against your data storage. Plans never shrink an existing disk.
Application backups contain the ProPR database, protected configuration and encryption keys, stored agent credentials, ProPR settings, and the queue state. They are compressed and encrypted on the VM before the remote copy. Repository clones, agent workspaces, and logs are excluded; clones come back from GitHub, and unpushed work in a workspace is not recoverable from a backup. The plan's backup retention is a rolling window while the instance runs. Restore is an operator-run procedure with no recovery-time commitment.
Each Hosted subscription gets its own VM, disk allocation, managed tunnel, and tenant identity. Tenant VMs cannot reach each other.
Provisioning, resizing, suspension, and deletion run from saved infrastructure plans. An operator approves each plan by its hash, and the approval expires. Deletion needs two different operators.
The GitHub account that subscribed becomes the instance administrator. Activation also waits for a health report from the new VM.
One fleet executor host applies approved plans to the Proxmox hosts. It cannot approve plans, verify backups, or request deletion.
You supply the AI subscriptions or API keys your agents use. Providers bill that usage directly; ProPR adds no token markup.
Offboarding drains running work, takes a final backup, and waits out the retention schedule agreed at activation before the destroy plan runs. The encrypted backups left after destruction are kept for 60 days, a window separate from the plan's rolling retention. Purging them afterwards needs two operator approvals.
Repository clones, prompts, task history, logs, settings, and the AI provider credentials you configure live on the Hosted VM. The ProPR team operates that VM and can access it to run, secure, back up, and restore the service. Review artifacts land in GitHub.
With self-hosting, the same data stays on infrastructure you operate, and the ProPR team has no access to it. Connect remains optional for webhook routing and tunnels. Compare both data boundaries.
Pick Solo, Team, or Scale and the number of developer seats.
Sign in to Connect with GitHub and select the GitHub App installation the instance serves. An organization installation needs an organization admin.
Connect reserves an instance slot before it opens checkout. When every slot is taken, Connect shows Hosted as sold out and starts no charge.
Payment sets the plan entitlement. The fleet then provisions your VM through an operator-approved plan, and Connect shows each lifecycle step.
Sign in to your ProPR instance with the GitHub account that subscribed and claim administrator access. Then add your AI provider subscriptions or API keys in the instance settings. Those credentials stay on your VM and never reach Connect.
Each subscription gets its own VM, so Hosted capacity is limited. Connect checkout shows whether a new instance is available right now; when it shows sold out, check again later.