Start free. Flat plans with an AI budget included.

Yes, out of the box. Connect a personal or group access token and you are set. A custom API endpoint is only needed for self-managed instances.

One run is one agent execution: a labeled issue, a retriggered task, or one step of an orchestrator pipeline. Static analysis on its own does not consume runs. A run that fails early still counts, because the sandbox was provisioned; a run that never starts (for example, over quota) does not. The same answer holds for money: a run that fails, times out, or produces a merge request you throw away still spends what it spent, on the FragonForge AI budget and on your own key alike, because the model calls it made were real. Nobody can refund tokens that were already generated, so we do not pretend otherwise. A run that never starts spends nothing.

Every user account in your FragonForge organization uses a seat, regardless of role: someone who only reviews the agent's merge requests or reads run history counts the same as someone who triggers runs. Free and Solo include 1 seat, Team 10, Business 30. The limit is enforced when a user is added or accepts an invitation, so you cannot overrun it by accident.

Only if you ask it to: runs happen when you label an issue, trigger one from the API or CLI, or start a scan, so the volume is yours to control. The quality gate pre-checks every change against 33 tools, so review time goes into substance rather than lint. And review is deliberately the default gate: FragonForge merges only what you have explicitly allowed, which is nothing by default; opt-in autonomous merge covers only steps whose quality gate is green. Start with one repository and widen as your review capacity allows.

On Free, runs stop. The limit is a hard stop at 100; new runs are rejected with a clear error until the month resets or you upgrade. On Solo, Team, and Business, the limit is fair use: runs beyond the quota are parked and requeued instead of rejected. Parked runs start automatically when the new month begins (quotas reset per calendar month, UTC); nothing is silently dropped, and there are no overage charges.

Not forever, but they are honestly gated. Every run executes in an isolated sandbox on our own EU hardware, so limits follow real server capacity, not an autoscaling abstraction. As we add capacity and can see that higher limits hold up economically, we will raise them; we would rather run a service that is still here next year than a generous one that is not. No date promised, because we do not publish dates we cannot keep.

A curated set of models we operate for you, so a run works without you holding an API key at all. FragonForge AI is billed in euros against the monthly budget your plan includes: EUR 7.50 on Solo, EUR 25 on Team, EUR 75 on Business. The budget is an included allowance, not a second meter on the plan: what the plan sells is the run quota, and the budget is there so you can start without an API key and keep going on the models we operate. It is not sized to fund the full quota, and on your own key the quota is the only limit that binds, because those runs do not draw on the budget at all. Free has no budget and runs on your own key instead, with a ChatGPT seat covering the runs its holder starts. The budget resets at the start of each billing period and does not roll over. You see used and remaining under Settings, Billing and on the Cost page, and every run carries the euro amount it was billed. A repository's Effective Configuration states which path its runs use, so there is no surprise about what a given repo bills against.

Runs pause and wait for the next reset rather than failing or charging you something you did not agree to. A paused run shows the reset time and your options: wait, upgrade for a bigger budget, or switch that repository to your own API key. If you would rather runs keep going, an Owner can turn on overrun under Settings, Billing: runs continue past the budget up to a limit you choose, and the extra usage appears as a separate euro line on your normal invoice. Overrun is off by default, can be switched off again at any time with immediate effect, and becomes available after your first completed paid billing cycle. The standard self-serve limit is EUR 20 per month; you can request a higher one in the same place, and we usually answer within two to four business days.

Not on your own key: your key, your provider, your bill, and we never resell tokens or mark them up. FragonForge AI is the other path, where we buy the tokens and bill them to your included budget in euros. One honest caveat about visibility on the bring-your-own-key path: the cost dashboard meters the planning and indexing calls, not the agent's own model calls inside the sandbox (those are proxied for isolation and pass through unmetered). Set a spending limit in your provider's console and treat its billing page as the source of truth there. A dedicated key just for FragonForge keeps that page easy to read. Your organization settings also carry per-stage output-token ceilings for the planner (architecture, decomposition, judge, plan), which bound that stage without touching the provider console.

Email, at support@fragonstudios.com, and no promised response time. Enterprise is the only plan with a named commitment (business hours CET, next business day for critical issues, two business days otherwise), because it is the only one we staff for. We would rather write that down than let a paying customer assume a service level that is not there. Separately from support, the Terms commit to no uptime SLA on any plan today, and the status page says the same thing. Neither is where we intend to stay: a response-time commitment for the paid self-serve plans is on the roadmap as decided but undated, and an uptime number follows once we have the operating history to stand behind one.

Yes, in both directions, from the billing portal. An upgrade applies immediately and is prorated for the rest of the period. A downgrade is scheduled for the period boundary, so you keep the limits you paid for until then. Invoice history lives in the same portal.

Yes. Forge support is not a paid feature. Self-managed GitLab, GitHub Enterprise, Gitea, and Forgejo endpoints work on every plan, including Free.

You do it yourself in the billing portal, with no retention call and no negotiation. The cancellation takes effect at the end of the period you already paid for, so nothing is cut short. After that your plan drops to Free; if you are over the Free limits at that point (repositories or seats), existing data is kept but new runs are blocked until you are back within them.

Take it with you first: the billing page has a self-serve account data export that gives you your organization's data as JSON, with credentials and secrets deliberately left out. Then delete the organization from the app. That purges repository metadata, run history, the code index, stored keys and tokens, and your billing link. It is a GDPR-grade deletion, not a soft delete.

No. Runs execute on our infrastructure in the EU, in isolated per-run sandboxes. Your forge and your LLM provider are the only external systems a run talks to.

A run has a hard wall-clock limit that depends on the plan: 15 minutes on Free, 30 minutes on Solo, 60 minutes on Team, and 120 minutes on Business. Enterprise limits are not a fixed row; every one of them, run length included, is set per agreement. A run that hits the limit is stopped and counts like any other run.

Yes. Disconnect the repository and connect a different one; the slot is freed immediately.

Not today. FragonForge is a hosted service and connects to your forge's API from our infrastructure, so the instance must be reachable from outside your network. If your GitLab or GHE lives strictly inside a VPN, FragonForge cannot connect to it. The exception is Enterprise, where a deployment on your own hardware can be arranged by agreement (none is live yet, and we say so). Tell us if this is your setup; it moves the priority.