Pay for refactor clarity,
not per developer.

Because not every engineer opens the dashboard daily. Per-seat pricing punishes teams for inviting the people who most need visibility — EMs, PMs, tech leads. You pay for the org; viewers are always free. Swarmia and LinearB charge $39+ per seat. We don't.

In Standard mode (default), we store feature names and metrics as metadata — the same trust model Linear uses for your ticket titles. Code structure (line ranges, descriptions, dependency graphs) is encrypted at rest. In Private mode, even feature names and file paths are encrypted with your org key — our servers see only opaque IDs and counts. PR comments in Private mode are rendered by your CI runner, MCP becomes local-only, and Slack notifications switch to link-only stubs (or you run a small self-hosted notifier we provide as a Docker image). Trade-off: more privacy, more setup.

Never. The scanner processes code in memory only and discards it after analyzing. What we store is derived metadata — feature names, file paths, metrics — and which fields are encrypted depends on your privacy mode. In Sovereign mode (Enterprise) we don't even touch your code in memory; Faultlines runs inside your VPC.

None. We cover the LLM cost so open-source maintainers can map their repos for free. Capped at 5 public repos and 500 scans/month — enough for a healthy project. If you maintain something massive (kubernetes-scale), email us and we'll figure out a sponsored plan.

Two operations consume your monthly scan budget differently. **Incremental scans** (on push, on PR sync) are unlimited on every paid tier — they cost cents in LLM tokens and we cover them. **Full re-scans** (initial scan, on-demand re-analysis, weekly drift check) draw from your monthly cap (30 on Team, 100 on Business, unlimited on Enterprise). The dashboard shows running usage and projected monthly burn.

Boost packs are available on every paid plan and never expire: $30 for 200 scans, $120 for 1,000 scans, $450 for 5,000 scans. No auto-renewal — you choose when to add. Boost scans persist in your account until used. We email you at 80% and 100% of your monthly cap, with a 48-hour grace period before pausing so a Friday-evening merge never breaks your CI.

Honest answer: at scale you might. But the math behind "saved with BYO" is real only for very heavy users (300+ full re-scans/mo on a 5M LOC monorepo) — that's the Enterprise band where we'll negotiate BYO + custom pricing anyway. For everyone else, included LLM is cheaper than the time to set up a separate Anthropic / OpenAI billing relationship. If you genuinely want to BYO key for compliance or to capture your enterprise discount, email hello@faultlines.dev and we'll set up a custom Business arrangement.

Faultlines reads your Sentry stack-trace filenames and PostHog page URLs, then matches them against the scan that was active when the event occurred (using Sentry's release + commit SHA, or PostHog's $app_version env var). Errors and pageviews appear in the dashboard attributed to specific features and flows — no SDK changes, no manual tagging, no code instrumentation. Just connect both via OAuth or API key on the Integrations page.

Yes. Upgrade to Business at any time and flip the toggle in Org Settings → Privacy. Existing scans are re-encrypted with your org key over the next sync window (~5 minutes for typical sizes). PR comment generation switches to CI-based — we provide the GitHub Action and Slack notifier. Downgrade is also supported.

Yes. The CLI and MCP server are open source and will always be free. You bring your own LLM key (Anthropic, OpenAI, or local Ollama) and pay your provider directly. There are no file limits, no scan credits, no cloud storage. Paid tiers add the hosted dashboard, shared feature maps, PR automation, and runtime integrations — not the core analysis.

Yes, on Enterprise. Self-hosted Faultlines runs in your VPC with SAML SSO, custom data retention, and dedicated support. Bring your own LLM key and you control the entire data path. Email hello@faultlines.dev to start a security questionnaire.