NEWBravos is availableCreate your account

What you buy here@asset <path>Declares a named asset: a component, a service or a data store.@asset in the GuardLink reference

Bravos is free to run. A licence takes it to every machine and pull request.

Bravos Free costs nothing and needs no account: the research loop, in one repository. A licence adds the pre-PR gate and autopilot, CI and nightly workflows, scheduled upkeep, multi-repository workspaces, campaigns, SAST and the code graph on your developers’ machines, and GST tax invoices. bravos account activate starts a trial. The trial is the full Team bundle for 30 days. No card.

Buy Bravos Solo for one developer, or Bravos Team per seat from 2 seats, online and billed monthly or annually. Sign in at account.bugb.io to see prices and buy.

Bravos licences

Bravos Free

Free · no account, no licence

For one person, researching one repository.

npm install -g @bugb-technologies/bravos
Install guide (opens in a new tab)
  • The research loop: context, annotate, templatize, env, prove, recheck and report
  • bravos watch, a live view beside your agent
  • One repository

Bravos Solo

Researcher

1 seat · billed monthly or annually

For one person who wants the Team bundle.

Everything in Bravos Free, plus

  • The pre-PR gate and autopilot
  • CI and nightly workflows
  • Workspaces, campaigns and scheduled upkeep
  • SAST and the code graph
  • GST tax invoices

Bought online. No repository limit. Cancel any time from your account.

Bravos Team

Teams

Per seat, from 2 seats · billed monthly or annually

For a team sharing one licence and one invoice.

Everything in Bravos Solo, plus

  • Invite and remove your people
  • Owner, admin and member roles
  • One licence and one invoice for the team

Bought online. From 2 seats, you choose how many. No repository limit.

Bravos Enterprise

Contact sales · terms of your own

For procurement, your own terms, or an estate with no internet.

Everything in Bravos Team, plus

  • Seats and term on your contract
  • Air-gapped licence on a longer term
  • Your security review and a DPA
  • Single sign-on (SAML)

No repository limit.

Try it first
bravos account activate starts a trial on your machine. The trial is the full Team bundle for 30 days. No card.
Seats
Bravos Team starts at 2 seats. You can buy up to 50 seats online on monthly billing, or 6 on annual billing. For more, send a bulk request from your account and we reply within 24 hours.
Switching plans
Already subscribed? Switch between Bravos Solo and Bravos Team, or between monthly and annual, on account.bugb.io/pricing. Changes start with your next billing period.
Invoices and tax
Every payment gets a GST tax invoice from Bugb, with your GSTIN on it when you add it to your billing details. Prices are before GST, which is added when you pay, and everything is charged in Indian rupees.

Not sure what a seat is or how a licence renews? How licensing works (opens in a new tab)

From sign-in to a licensed machine@flows <source> -> <target> [via <mechanism>]Documents data moving from one component to another, and what carries it.@flows in the GuardLink reference

How buying works

  1. Create your account

    Sign in at account.bugb.io with Google, GitHub or an email link. Your first sign-in creates the account. There is no password.

  2. Add your organisation and billing details

    Name your organisation and say who to invoice. Add your GSTIN to your billing details and it appears on your GST tax invoices.

  3. Buy online, per developer or per seat

    Buy Bravos Solo for one developer, or Bravos Team from 2 seats, by card from the account, billed monthly or annually. For more seats than you can buy online, or an Enterprise contract, we set it up with you.

  4. Licence your machines

    Your licence is issued as soon as the plan is active. Each developer runs bravos account activate, or imports the licence file, then bravos setup fetches SAST and the code graph. For CI, add the licence as the BRAVOS_LICENCE repository secret. Invite your team from the account.

Runs on your agent's subscription@flows <source> -> <target> [via <mechanism>]Documents data moving from one component to another, and what carries it.@flows in the GuardLink reference

What a pentest uses

Bugb does not charge per pentest. Bravos runs inside the coding agent you already use, so its AI work counts against that agent's subscription, like the rest of your coding. Replaying tests you already approved uses no AI at all. How much new testing it does depends on the level you choose.

See how much of your plan it uses

Pentest levels and how much AI each one uses
LevelWhat it doesAI use
LowNo new tests. Replays the tests you already approved.No AI use
MediumdefaultAlso tests findings that have no test yet, up to 8 per run. It asks first, unless you said yes to continuous pentesting.Up to 8 findings tested per run
HighUp to 32 findings per run, and chains them into multi-step attacks.Up to 32 findings tested per run, plus chaining
MaxEvery testable class of exposure. It is approved in advance, so it does not ask.Grows with the size of your threat model
What it runs on
Your coding agent's subscription: Claude Code, Codex, Cursor, Windsurf, opencode or another. Bravos calls no AI model of its own, and Bugb adds no charge per pentest.
Continuous pentesting
If you answer yes in bravos start, Bravos tests again whenever changed code adds or changes an exposure. How much of your plan that uses follows how often it happens, at the level you set.
Scheduled runs
A scheduled run in GitHub Actions starts your agent there, signed in with the key or token you store as a repository secret. A weekly run uses about a fifth of a weekday one.
Keeping it at zero
Answer no to continuous pentesting, or set the level to low. Bravos still replays the tests you approved, which uses no AI.

Set the level by telling your agent (“pentest at high”), or add the line bravos effort: high to AGENTS.md or CLAUDE.md.

Services and assessments

When you want us to do the work: a threat model, a pentest, or the whole loop running in your CI/CD. We discuss your repositories, targets and access before proposing a scope.

  • Threat modeling as code

    Make security context part of your repository, not a document that drifts away from it.

    • GuardLink code annotations
    • Assets, threats and security assumptions
    • A threat model your team can review in Git
    Talk to sales
  • Pentesting as code

    Turn pentest hypotheses into executable tests you can keep, inspect and replay.

    • Cert-X-Gen pentest templates
    • Evidence-backed test outcomes
    • Replay stored probes after a fix
    Talk to sales
  • Continuous threat modeling and pentesting

    Connect the whole round in CI/CD with Bravos. Keep the model, pentests and evidence connected as your code changes.

    • Annotation, environment and authentication orchestration
    • Pentest execution and vulnerability ledgers
    • Evidence-backed annotation write-back
    • Re-testing and CI gates based on your policy
    Talk to sales
  • Whitebox AI-agentic pentesting

    Focus on an assessment with source-code context to guide hypotheses and testing.

    • Repository-informed threat hypotheses
    • Authenticated application testing
    • Findings with evidence and code context
    Talk to sales
  • Blackbox AI-agentic pentesting

    Focus on an assessment from the outside, without requiring source-code access.

    • Testing against an agreed target scope
    • Externally observable application behavior
    • Evidence and remediation guidance
    Talk to sales

Our open-source tools stay free on their own: GuardLink and Cert-X-Gen. Every engagement follows our responsible disclosure pledge.