Pricing that scales
with your team
Full-platform software engineering development — free to start, transparent per-seat pricing for teams, and custom enterprise plans with governance and scale. No hidden fees, no annual lock-in.
At $100/seat/mo, Pro works out to about $0.06 per shipped acceptance criterion at our published benchmark rate (286 acceptance criteria shipped across 78 verified tasks in a 5-day window, last measured August 2026). Benchmarks methodology documents how the figure is measured.
- 2 agents (BA + Coding)
- All core features included
- Full product experience, scale-limited
- Email support
- BYOK inference
- 2 agents per user (BA + Coding)
- Fully managed
- Email support
- BYOK inference
- 1-way integrations (Jira → driftless)
- Additional agent types (paid add-on)
- Custom skills (paid add-on)
- Unlimited, org-scoped agents
- Premium support — dedicated CSM + white-glove onboarding
- SSO: SAML/OIDC (Okta, AD, Entra)
- SCIM user provisioning
- Audit logs with export
- 2-way integrations (Jira, GitHub)
- Custom SOUL tuning & custom skills included
- SLA guarantees & data residency
- Single-tenant option ($100K–$500K/yr)
- Custom RAG / knowledge base
Feature comparison
Everything included at every tier. Enterprise adds governance, security, and customization for organizations that need it.
| Feature | Free | Pro | Enterprise |
|---|---|---|---|
| Core Platform | |||
| Shaping (PRD / tech spec pairing) | |||
| Task board (Kanban workflow) | |||
| Agents | |||
| Analytics & cost tracking | |||
| File uploads on tasks | |||
| API / MCP access | |||
| Agents | |||
| Agent count | 2 (BA + Coding) | 2 per user | Unlimited, org-scoped |
| Agent types | BA + Coding | BA + Coding | All types included |
| Custom SOUL tuning | — | — | |
| Custom skills | — | Add-on | |
| Support | |||
| Support level | Premium (dedicated CSM) | ||
| Onboarding | Self-serve | Self-serve | White-glove |
| Dedicated CSM | — | — | |
| Security & Enterprise | |||
| SSO (SAML/OIDC) | — | — | |
| SCIM user provisioning | — | — | |
| Audit logs with export | — | — | |
| SLA guarantees | — | — | |
| Data residency | — | — | |
| Single-tenant deployment | — | — | |
| Custom RAG / knowledge base | — | — | |
| Integrations | |||
| Integration direction | None | 1-way (Jira → driftless) | 2-way (Jira, GitHub) |
| Supported platforms | — | Jira | Jira, GitHub, + more |
Pricing FAQ
Answers to the questions evaluators ask before picking a tier. Claims are grounded in shipped functionality; comparison links point to the fact-checked /vs pages rather than duplicating them.
How is driftless different from a project tracker?
A tracker records and routes work. driftless runs the delivery cycle itself: the BA agent shapes paired PRDs and tech specs, spec gates run as platform workflow, agents execute against task directives, and shipped work is validated against acceptance criteria. The methodology replaces ceremony with software-enforced structure, and the published benchmark documents the delivery rate at /docs/benchmarks.
Do I need to bring my own LLM keys?
Yes. Inference runs on your own provider keys: Anthropic, OpenAI, and OpenRouter are preset providers, and you can point driftless at any OpenAI-compatible endpoint you control. Keys are encrypted at rest in a per-user vault, proxied server-side, never exposed to the browser, and revocable instantly. Tasks, initiatives, PRDs, and tech specs are exportable at any time. See BYOK and key sovereignty.
How does driftless compare to GitHub Spec Kit or Kiro?
The fact-checked /vs pages are the single source of truth for those comparisons: read driftless vs Spec Kit and driftless vs Kiro for the verified, current facts.
What happens when a spec gate fails?
Work is blocked in place. The gate returns specific, actionable findings pointing to the exact section and the exact problem, and the task waits for a person to decide. Nothing auto-closes, failed work is never marked done, and dependency chains prevent agents from starting blocked work.
Does driftless work with the coding agents we already run?
Yes. driftless speaks MCP and ACP plus a documented REST API, so any MCP-compatible client connects without replacing your toolchain. The orchestrator runs on your machine, polls the task board, and dispatches your coding agent against your repos with your keys. You choose where the work executes: hosted agents run in the hosted runtime, your coding agent runs on your machine.
Is there a free tier?
Yes. Free is $0 for 1 user, self-managed, with two agents (BA + Coding), all core features included, email support, and BYOK inference. No credit card required.
Drag the slider to see your monthly cost with Pro volume pricing. When you're ready, submit your quote and our team will reach out.
Something went wrong. Please try again or email us at sales@godriftless.ai.
Tell us which enterprise capabilities matter to your organization. Select the features you need, share your contact details, and our enterprise team will reach out within 1 business day.
Everything you get
Not just task management — the entire Cascades methodology codified into the platform. Full observability, traceability, and auditable delivery, included at every tier.
Shape
BA agent converts conviction into structured, testable, sequenced work. PRDs define what and why; tech specs define how.
Gate
Structural and semantic validation before code is written. Every task has acceptance criteria; dependencies are linked; no orphaned work.
Ship
Orchestrated coding agents execute against context-engine directives — initiative details, task directives, acceptance criteria.
Verify
Automated validation that shipped work meets the acceptance criteria defined during shaping.
Cascades Methodology
The entire Shape → Gate → Ship → Verify cycle, codified into the platform. Sequential phases, tight loops, small drops, zero ceremony tax.
- Shape — PRD + tech spec pairing
- Gate — structural & semantic validation
- Ship — orchestrated coding agents
- Verify — automated AC validation
Task Board & Project Management
Kanban board with full status workflow, acceptance criteria on every task, and nine dependency link types for cross-project visibility.
- Kanban: backlog → done
- Acceptance criteria in Given/When/Then
- 9 dependency link types
- Task type tagging
- Full audit trail via comments
Agent System
Custom BA Agent with tuned SOUL, agent memory, and a skills system — purpose-built for requirements shaping, not a generic chatbot.
- BA Agent with tuned SOUL
- Skills: bug-intake, NFR, shaping, TDD
- Agent memory — persistent, topic summaries
- Important info — durable, never summarized
- Multi-agent support (BA + coding)
Observability & Traceability
Full track-and-trace: every requirement traces to a task, every task traces to a requirement. Comment threads serve as decision logs.
- Requirement-to-task traceability
- Initiative-level spec pairing
- Comment threads as decision logs
- Task status as a signal, not a vanity metric
Analytics & Cost
Org and project dashboards with per-model cost breakdowns, delivery metrics, and agent LLM usage tracking.
- Org & project dashboards
- Per-model cost breakdown + cache hit rate
- Delivery metrics (ACs, cycle time)
- Agent LLM usage tracking
Integrations & API
API and MCP access live today, plus file uploads on tasks — images, video, audio, and documents.
- API access (live)
- MCP integration (live)
- File uploads (images, video, audio, PDFs)
- Expanding to text, CSV, Excel, Word
How does driftless compare?
Answers to the comparison question evaluators ask most. Each links to the fact-checked /vs page, which is the single source of truth for these facts.
driftless vs Kiro
Kiro checks requirements for contradictions and gaps inside the IDE and prices by credits per user. driftless enforces spec gates as platform workflow and prices per seat with BYOK inference. Read the comparison
driftless vs drft
drft is a governance layer around coding agents you already run, connected through MCP. driftless includes the persistent agents, shaping, gates, tracking, and verification in one platform. Read the comparison
driftless vs Spec Kit
Spec Kit is a free, MIT-licensed toolkit your coding agent executes in your repo, with opt-in quality gates. driftless is the managed platform with a BA agent and platform-enforced spec gates. Read the comparison
driftless