Simple pricing.
Your keys, your model bill.
One flat fee for the platform. Model usage runs on your own Anthropic key, at cost — no metering, no markup, no surprises.
Free
Work 25 real tickets a month. No card.
- 25 runs / month
- 1 concurrent run · 1 Lane
- GitHub Issues & Shortcut ticketing
- All repo hosts — GitHub, Bitbucket, Azure
- Optimize (repo map + guide) included
- Multi-repo, notifications — from Solo
Solo
For one developer putting a real backlog to work.
- 400 runs / month
- 2 concurrent runs · 3 Lanes
- Multi-repo Lanes
- Slack & Discord notifications
- Coding Standards module
Pro
The daily driver — more volume, more parallelism, first in line for what's new.
- 1,500 runs / month
- 5 concurrent runs · 10 Lanes · 3 seats
- Jira & Azure Boards ticketing
- Cross-ticket history & Agent Tuning
- Everything in Solo
- Early access: new agents & beta features
Team
For teams shipping through Tenbi together.
- 5,000 runs / month (fair use)
- 15 concurrent runs · unlimited Lanes
- 10 seats included, $10 / seat after
- Priority queueing
- Everything in Pro
- Admin panel, audit log & per-seat analytics
- Priority support
Enterprise
Dedicated machines, no queueing, custom quotas, DPA and custom terms. Built around your org.
Why is this so much cheaper than a bundled agent?
Because we don't resell model tokens. Tenbi is BYO-key: every run uses your own Anthropic API key, so model usage bills to your account at Anthropic's prices — no markup, no credits, no meter. Real completed runs measure by ticket size: small fixes $2–$4 in Claude usage; feature-sized tickets $5–$20 (largest measured ≈ $17) — billed by Anthropic to you, and every run reports its exact cost live. The plan pays for the orchestration platform: the agent pipeline, guardrails, multi-repo workspaces, and observability. How we store your keys →
What a ticket actually costs
Every agent vendor charges differently. Here's the pricing structure, side by side.
| Platform fee | Usage model | A ticket's model cost | |
|---|---|---|---|
| Tenbi | $0–199/mo flat | Your own Anthropic API key — at Anthropic's prices, no markup | $2–$20 measured, by ticket size, billed by Anthropic to you |
| Devin | from $20/mo | ~$2.25 per ACU (an ACU ≈ 15 minutes of agent work) | varies with ACUs consumed; per-task cost not published |
| GitHub Copilot coding agent | from $10/mo | dollar-denominated credits per plan tier | varies with credits consumed; per-task cost not published |
| Atlassian Rovo Dev | $20/dev/mo | 2,000 credits included, then $0.01 per credit | varies with credits consumed; per-task cost not published |
Competitor rates as published on their public pricing pages, July 2026. Our number is measured across real completed runs in the beta and will be re-baselined as usage grows — heavier repositories and tickets run higher.
Compare plans
| Free | Solo | Pro | Team | Enterprise | |
|---|---|---|---|---|---|
| Price | $0 | $19/mo | $39/mo | $199/mo | Custom |
| Runs per month | 25 | 400 | 1,500 | 5,000* | Unlimited |
| Concurrent runs | 1 | 2 | 5 | 15 | Custom |
| Enabled Lanes | 1 | 3 | 10 | Unlimited | Unlimited |
| Seats | 1 | 1 | 3 | 10 + $10/seat | Custom |
| Repo hosts (GitHub / Bitbucket / Azure) | ✓ | ✓ | ✓ | ✓ | ✓ |
| Ticketing: GitHub Issues + Shortcut | ✓ | ✓ | ✓ | ✓ | ✓ |
| Ticketing: Jira + Azure Boards | — | — | ✓ | ✓ | ✓ |
| Multi-repo workspaces | — | ✓ | ✓ | ✓ | ✓ |
| Slack / Discord notifications | — | ✓ | ✓ | ✓ | ✓ |
| Optimize (repo map + guide) | ✓ | ✓ | ✓ | ✓ | ✓ |
| Coding Standards module | — | ✓ | ✓ | ✓ | ✓ |
| Agent Tuning module | — | — | ✓ | ✓ | ✓ |
| Cross-ticket history | — | — | ✓ | ✓ | ✓ |
| Early access (new agents, beta) | — | — | ✓ | ✓ | ✓ |
| Multi-user + permissions + analytics | — | — | — | ✓ | ✓ |
| Priority queueing | — | — | — | ✓ | Dedicated |
| Machines | Shared | Shared | Shared | Shared | Dedicated, no queue |
| Support | Community | Priority email | Priority | Dedicated + SLA | |
| Model costs | Always your own Anthropic API key — billed to you at Anthropic's prices, never marked up | ||||
* Team's 5,000/month is a published fair-use line (≈230 runs every workday) for unlimited normal team development — see Terms of Service and the FAQ.
Questions
What counts as a run?
One ticket worked end-to-end — picked up, planned, coded, reviewed, and delivered as pull requests. A multi-repo ticket that opens several PRs is still one run. Re-running a ticket counts as a new run. System-triggered runs — Optimize refreshes and indexer passes — don't count against your cap; we trigger those, so they're on us.
What happens when I hit my monthly run limit?
New tickets simply stop being picked up until your month rolls over (or you upgrade). Nothing fails, nothing is lost — runs in flight finish normally, and waiting tickets are picked up when capacity returns.
Which AI models does Tenbi use?
Claude, exclusively. Tenbi is purpose-built around Anthropic's Claude models and agent tooling — not a lowest-common-denominator multi-model wrapper. You'll need an Anthropic API key; other model providers aren't supported.
Why do I also pay Anthropic?
Tenbi is BYO-key: the AI model runs on your own Anthropic API key, so you pay Anthropic directly at their prices — this is also the way Anthropic's terms require third-party products to work. We never mark up or resell tokens; your plan covers the orchestration platform only. Small fixes usually run a few dollars in Claude usage; feature-sized tickets typically run $5–$20 — and every run shows its exact cost live.
Can I use my Claude subscription instead of an API key?
No — Tenbi runs on an Anthropic API key. Anthropic's terms restrict subscription credentials to Anthropic's own products, and we build to their terms. API usage is billed by Anthropic at the same per-token prices you'd pay them directly.
What does "early-access pricing" mean exactly?
Pro is $39/mo during beta. At GA the list price becomes $59 for new customers — everyone who subscribed before the change keeps $39 for as long as they stay subscribed. No coupons, no countdown games; this and the annual discount are the only promotions we run.
Can I change plans later?
Yes — upgrades take effect immediately, downgrades at the end of the billing period. Your Lanes and history are unaffected either way.
Do unused runs roll over?
No — run allowances reset at the start of each billing month.
What does "fair use" mean on Team?
Unlimited means unmetered for normal team development, with a published line so nothing is hidden: 5,000 runs a month — about 230 every working day. Automated bulk pipelines that treat Tenbi like a CI farm fall outside fair use — see the Terms of Service. If you need that kind of scale, Enterprise gives you dedicated machines built for it.
Is there an SLA?
Not during beta. Enterprise plans carry dedicated support and SLA terms — contact us.