Run timeline
- run started
- analyzing codebase
- implementing changes
- running tests
- creating pull request
- awaiting review
Give Vif a GitHub issue. It runs the work in a scoped sandbox, opens the PR, and brings back the context reviewers need before the queue keeps growing.
See pricingvif workflow run
42 func StartDeviceFlow(ctx context.Context, client *oauth2.Config) (*DeviceFlow, error) {
43- // old implementation
44+ // device flow implementation
45+ df, err := client.DeviceAuth(ctx, scopes...)
46+ if err != nil {
47+ return nil, err
48+ }
49 return df, nil
}
How it works
Vif keeps the workflow inside GitHub, but makes each handoff explicit: what starts the run, where code executes, what lands in the pull request, and when your team takes back control.
Use the normal backlog item as the source of truth instead of opening a separate agent task.
vif label or comment /vif implement.The implementation happens in the workspace selected for your access, capacity, and isolation needs.
Reviewers get the artifact they already know how to inspect: branch, commits, tests, and summary.
Vif prepares the PR. Your engineers still decide what is acceptable, what changes, and what merges.
Vif is for engineering teams with more requested work than reviewer time. It takes a real GitHub issue, runs the implementation in a scoped workspace, and returns a pull request your team can inspect, request changes on, and merge.
Use Vif when the problem is already described in GitHub and the next useful artifact is a reviewed PR.
Start with Vif Cloud for standard GitHub work, or size a dedicated setup when private systems and team-specific tool rules matter.
The simulator estimates how many coding workflows can move at once, then flags when the setup should be quoted instead of self-served.
Vif takes a GitHub issue, runs your connected coding agent in an isolated workspace, and returns the work as a pull request.
Vif is not a coding editor. It manages issue intake, workspace isolation, the run, PR creation, review status, and follow-up work around GitHub.
Your connected agent account runs the work. Vif pricing covers orchestration, workspace capacity, queueing, and review handoff rather than reselling tokens.
Use a quoted runtime when work needs private systems, internal tools, custom MCP servers, team-specific access rules, compliance review, or unusual resources.
The short version: Vif does not replace your review process. It gives implementation work a scoped place to run, then brings the result back as a pull request your team already knows how to judge.
Vif uses the coding agent account your team connects. Bring your own agent means the work runs through your account, not a generic Vif-owned account.
Vif reads the issue, starts a scoped run, prepares the workspace, tracks the work, and returns the result as a normal pull request with notes and status.
Your team does. Vif prepares the PR and can handle follow-up work, but approval, requested changes, merge timing, and production responsibility stay with your engineers.
Vif is agent-agnostic by design and BYOA-first in operation. Hosted Vif supports OpenAI Codex through your connected account today; dedicated or custom setups can scope additional agent profiles as they become available for your team.
Yes. A workflow can choose the connected agent account, model, and effort level for each step, then combine that with approval rules, workspace boundaries, and allowed systems.
No. Workflow steps such as implementation, review, and fixes run inside isolated containers with the workspace boundary assigned to that step.
Managed hosted pricing is based on concurrent capacity slots, not token resale. Dedicated runtime and custom deployment setups are quoted from the access boundary, support, and resource requirements.
A slot is one issue-to-PR workflow running at the same time. Slot size controls the runtime resources available to that workflow, and extra compatible work queues until capacity is free.
Choose a dedicated or custom setup when the work needs private systems, internal documentation, custom tool connectors, stricter access rules, procurement, or security review.
No. Vif moves implementation work into a pull request. Your engineers still review, request changes, approve, merge, and own production responsibility.
Shared hosted pricing is for standard GitHub work and published capacity slots. Private systems, custom tool access, and security review move the request into a quoted dedicated or custom setup.
Use managed hosted slots when the workflow only needs GitHub, your connected agent account, and the default hosted rules.
Vif bills for orchestration and workspace capacity. Model usage remains with the account your team connects.
Private networks, internal tools, custom MCP servers, compliance, or unusual resources are scoped before pricing is finalized.
Start with the boundary your team is comfortable with. Hosted is fastest when it fits; private systems need a quoted setup.
Best fit when your team only needs standard GitHub issue-to-PR work and the hosted boundary is acceptable.
Use the managed hosted runtime when Vif does not need private systems, custom access rules, or procurement review. Buy concurrent slots, install the GitHub App, connect your agent account, and Vif queues work when all selected slots are busy.
Custom capacities are not billed from this simulator. They are quoted with workspace isolation, support, and resource requirements.
SSO, audit logs, role-based access, and approval flows can be part of a custom deployment.
Tools, secrets, and sandboxes stay inside the workspace boundary recommended for your team.
Run in your own cloud or use Vif Cloud. Your code, your choice.