Activepieces starts at $0 on Free, while Plus costs $16 a month billed yearly and Team costs $166 a month billed yearly, verified 6 September 2026. Plus includes 10,000 monthly credits and up to five users; Team includes 50,000 monthly credits, 25 users, SSO, and standard roles. Ultimate is custom-priced, and Embed has no published price. These are flat plan prices, not per-user starting prices.
The important qualification: one normal flow run costs one credit, but AI adds credits on top. Plus and Team charge $0.007 per extra credit. Free self-hosted Community Edition avoids a software subscription for the automation core, but does not include the whole paid collaboration and governance suite—or eliminate infrastructure and maintenance costs.
For most small teams, price Plus first, then check whether governance requires Team or Ultimate. For self-hosting, compare operating cost and required features, not just the license price. All vendor prices and limits below refer to the official pricing page, checked 6 September 2026. Dollar calculations use USD budgeting assumptions, exclude tax, and use annual-billing monthly equivalents where applicable.
AI-generated editorial illustration of the cost components; not an Activepieces interface or invoice.
Activepieces pricing at a glance: Free, Plus, Team, Ultimate, and Embed
Plus is the low-cost paid starting point; Team is the first published tier for SSO and standard roles. A small headcount does not automatically mean the smallest plan meets your requirements.
| Plan | Published price, checked 6 September 2026 | Credits and users | Main buying boundary |
|---|---|---|---|
| Free | $0; no card or time limit | 100 credits per day; one user | Lightweight exploration; excess runs wait for the next daily refresh |
| Plus | $16/month, billed yearly; $192 annual base equivalent | 10,000 credits/month; up to five users | BYOK available; $0.007/extra credit; community human support |
| Team | $166/month, billed yearly; $1,992 annual base equivalent | 50,000 credits/month; 25 users | Unlimited projects, global connections, SSO, standard roles, email support; $0.007/extra credit |
| Ultimate | Custom | Negotiated yearly credit pool; unlimited users | Dedicated workers/support, execution priority, and enterprise controls |
| Embed | No public price; sales-led evaluation | Obtain a written entitlement and usage quote | White-label builder, AI agents, and integrations inside your own product |
The four main plans list unlimited flows. That means the number of automations you build is not the same as the amount of usage included. Their product matrix also lists Flows, Agents and Chat, Tables, MCPs, team analytics, and API/SDK access; those labels do not establish unlimited storage, retention, throughput, or model usage. Ask for those limits if they affect your design.
Ultimate adds SCIM, custom RBAC, releases/Git Sync, audit logs, secret managers, event streaming, private pieces, project credit reporting, branding, platform governance, custom terms, and per-project execution priority. Do not assume Team includes those controls merely because it includes SSO.
The pricing page offers an annual/monthly toggle. The verified prices above are the annual-billing view; they are not a claim that month-to-month checkout costs the same. Confirm billing cadence, taxes, cancellation terms, extra-seat availability, and any negotiated terms before purchase. The FAQ mentions prorated upgrades and added seats but does not publish an extra-seat rate.
How Activepieces credits work, including AI and overage
Count whole flow runs first, then add AI steps and agentic actions. A long non-AI workflow can therefore consume fewer credits than a short AI-heavy one.
The official pricing FAQ, verified 6 September 2026, gives these rules:
| Usage component | Activepieces credits |
|---|---|
| Normal flow run, regardless of step count | 1 per run |
| AI step using a fast model | 2 additional credits |
| AI step using a smart model | 10 additional credits |
| AI step using a frontier model | 20 additional credits |
| Agentic action | 1 each |
| AI step using your own provider key, on Plus and up | 1 additional credit, plus external provider charges |
For a predictable workflow using vendor-supplied models:
Monthly credits = flow runs + (2 × fast AI steps) + (10 × smart AI steps) + (20 × frontier AI steps) + agentic actions.
Use actual executed counts across the month, not just nodes drawn in the editor. Conditional branches and agents can vary. The scenarios below assume no retries or replays; verify their treatment, unsuccessful calls, tests, and any nested-flow accounting with Activepieces before using the formula as a purchase commitment.
AI-generated explanation of the published credit rules, checked 6 September 2026. Model tiers are billing categories, not a guarantee of task quality.
Overage is $7 per additional 1,000 credits
Plus and Team continue running above allowance and charge $0.007 per extra credit, as checked 6 September 2026. The normalized monthly-equivalent calculation is:
Plan base + max(0, monthly credits − included credits) × $0.007.
Free behaves differently: flows stay enabled, but over-quota runs wait until the next refresh. One hundred daily credits is not a bankable 3,000-credit monthly allowance. A 400-credit Monday cannot borrow unused credits from Tuesday. That makes Free a poor fit when delayed processing would lose leads or interrupt operations.
Bring your own key moves the model bill; it does not erase it
BYOK on Plus and higher reduces each AI step to one Activepieces credit while you pay the model provider separately. It is unavailable on Free under the published plan matrix, checked 6 September 2026.
That can be valuable for heavier AI usage, but calculate provider tokens, tools, and other applicable charges. Add the operational work of controlling access, rotating keys, setting budgets, and reconciling two bills. BYOK is a billing option, not proof that data stays on your own servers or that every model is suitable for sensitive information.
Four realistic Activepieces cost scenarios
Usage, governance, and staff time can each dominate the price. These are illustrative budgets, not measured workloads, vendor infrastructure quotes, or guaranteed outcomes. All plan inputs use the verified 6 September 2026 annual-billing prices.
1. Non-AI operations: 20,000 monthly runs
Assume one owner automates CRM updates and notifications: 20,000 runs, eight ordinary steps per run, no AI or agentic actions. The workload consumes 20,000 credits, not 160,000.
Plus is $16 + (10,000 × $0.007) = $86/month equivalent. Team is $166/month equivalent, with no overage at this volume. If the operator spends two hours monthly checking failures and maintaining workflows at an assumed $60/hour loaded cost, the recurring economic totals become $206 on Plus or $286 on Team. Connected application subscriptions are additional to both.
For usage alone, Plus and Team reach parity at approximately 31,429 monthly credits, while Team remains within its 50,000 allowance: $16 + (credits − 10,000) × $0.007 = $166. Above 50,000, both use the same published overage rate; Team's extra included credits leave it $130/month cheaper on this simplified usage calculation. Features or custom arrangements can change the decision.
2. AI-heavy triage: 10,000 runs with one smart AI step each
Assume every run makes exactly one smart-model classification, with no agentic actions. Usage is 10,000 × (1 + 10) = 110,000 credits.
| Route | Monthly-equivalent software calculation | Software total |
|---|---|---|
| Plus, vendor model | $16 + 100,000 × $0.007 | $716 |
| Team, vendor model | $166 + 60,000 × $0.007 | $586 |
| Plus, BYOK | $16 + 10,000 × $0.007, plus provider usage | $86 + provider bill |
With BYOK, the same count of AI steps uses 20,000 Activepieces credits. Assume—not quote—a $100 provider bill. That makes the BYOK software budget $186. Add three hours of monitoring and quality review at $60/hour, and the recurring economic totals are $896 for Plus/vendor model, $766 for Team/vendor model, or $366 for Plus/BYOK.
Against Team's $586 vendor-model bill, Plus/BYOK remains cheaper on software cost if its external provider bill is below $500/month. This is arithmetic, not an assertion of equivalent model quality or eligibility for the same workflow. Agentic actions, additional model calls, and review failures increase cost; a fixed single-call classification does not represent an open-ended agent.
3. Five users: headcount fits Plus, SSO does not
Assume five staff share 8,000 non-AI runs a month. Plus covers the published user limit and credit allowance for $16/month equivalent total, not $16 multiplied by five.
If company policy requires SSO and standard roles, the published floor is Team at $166/month equivalent, despite having only five users. Add an assumed two admin hours at $60/hour: $136 recurring economic cost on Plus versus $286 on Team. An assumed eight-hour initial workflow setup adds $480 once to either model; identity configuration may require more work on Team.
If the requirement is audit logs, SCIM, custom roles, or controlled releases, get an Ultimate quote. Do not present the Team price as a complete answer to an enterprise-control requirement. For a sixth user without those requirements, ask about seat expansion terms rather than inventing a published add-on price.
4. Self-hosted Community Edition: a non-AI production budget
Assume a competent operator needs only the automation core, not the paid collaboration or governance layer. A planning budget might be:
| Assumed recurring cost | Monthly budget |
|---|---|
| Compute and storage | $30 |
| Backups | $10 |
| Monitoring | $10 |
| Three operator hours at $60/hour | $180 |
| Recurring total | $230 |
Assume ten initial deployment and recovery-testing hours at $60/hour: $600 setup. The first-year model is $600 + 12 × $230 = $3,360, or $280/month averaged across year one. A four-hour incident adds $240. External AI, paid business apps, commercial edition fees, and redundancy requirements are outside this non-AI example and must be added when applicable.
These infrastructure figures are budget placeholders for calculation—not promised hosting prices or a sizing recommendation. Traffic, databases, history retention, resilience, and operator skill determine the real number. Cloud also needs workflow maintenance, but Community Edition additionally needs someone accountable for the host and recovery.
What free self-hosted Community Edition excludes
Community Edition is the automation core, not “Ultimate on your own server.” The official pricing FAQ explicitly separates hosting location from product entitlement.
The page says its plans can run in the cloud or on-prem at the same plan price. Separately, it explains that paying adds access control, change visibility, team/client separation, safer releases, Agents and Chat, and vendor support beyond a forum. Do not transfer the Free Cloud product list to free Community Edition: “Free” on the hosted pricing table and the free self-hosted edition are different purchasing contexts.
AI-generated editorial illustration. Community Edition removes the core software fee, not the operating responsibilities.
For installation, start with the official install overview; the documented recommended self-host route uses Docker Compose with PostgreSQL and Redis. A working container is only the beginning of production ownership. Assign updates, TLS and network protection, backup retention, restore drills, credential rotation, alerting, and incident cover. Set a recovery objective before critical workflows depend on the system.
Self-hosting fits a business that has a qualified owner and a specific control or deployment reason. It is a poor bargain when the only owner is already overloaded, nobody can restore the database, or a commercial entitlement is needed anyway. Confirm the current edition feature matrix against the deployed version before migration.
Which Activepieces plan fits your business?
Choose the cheapest plan that clears both your operational requirements and your expected usage—not simply the lowest sticker price.
- Free: one person exploring lightweight workflows who can tolerate the daily limit and delayed runs. Avoid it for unattended, time-sensitive burst processing.
- Plus: up to five users with modest credit needs, no Team-level identity requirements, and acceptable community support. Check overage once usage approaches the calculated Team crossover.
- Team: buyers needing SSO, standard roles, project separation, global connections, email support, or enough credits to justify the larger allowance. It is not the audit-log or custom-RBAC tier.
- Ultimate: buyers whose mandatory controls include SCIM, custom roles, audit logs, release management, dedicated workers, or enterprise support. Get a written scope and yearly-pool agreement.
- Community Edition: technically owned core automation where missing paid features are acceptable. Budget infrastructure and labor before comparing it with Cloud.
- Embed: software vendors adding automation inside their own product. Evaluate tenant isolation, branding, SDK work, support responsibility, and commercial terms; it is not merely another internal SMB user plan.
These fit boundaries use the published matrix checked 6 September 2026. A plan's “team analytics” label is not a substitute for confirming the execution history, reporting exports, and retention your operator actually needs.
How to compare the cost with n8n, Make, and Zapier
Compare a matched workload, not nominal credit or task counts across vendors. Their billing units are not interchangeable.
n8n is a useful execution-based comparison for long workflows; its self-hosted route still needs technical ownership. Make is worth evaluating for visual branching, but model the billable module activity and AI behavior rather than treating a scenario as one Activepieces credit. Zapier is a useful ease-of-operation and connector comparison, but successful billable actions can multiply across a multi-step workflow. Connector availability alone does not guarantee the exact action, permission scope, or error handling you need.
For switching motives and alternatives, use our n8n alternatives guide. For a broader shortlist, read the small-business automation buyer guide. This pricing page does not rank competitors by an unverified current subscription price.
Run the same sample through each finalist: trigger volume, conditional paths, required actions, retries, AI calls, human approvals, and support needs. A tool costing less per unit can still cost more to operate if staff spend longer fixing it.
What to verify before committing
Build one representative workflow and reconcile its usage before an annual commitment. That is a more useful next step than extrapolating from a simple demo.
- List required users, projects, identity controls, support, and change-management features. Establish the plan floor first.
- Count expected runs, peak-day demand, AI steps by tier, and agentic actions. Include a separately labeled burst/retry allowance.
- Confirm exact integration actions, API rate limits, credential permissions, execution limits, and history retention. Unlimited flows is not unlimited capacity.
- Compare vendor-model billing with BYOK using actual representative token lengths and output quality. Keep human review costs visible.
- Ask about usage alerts, spend controls, failed-run accounting, cancellation, and seat expansion. Do not assume overage has a hard budget cap.
- For self-hosting, assign an owner and test a restore. For governance-led purchases, obtain a written feature and support schedule.
The pricing FAQ checked 6 September 2026 says Activepieces' SOC 2 program is in progress. Do not treat it as currently certified on that evidence. If procurement requires assurance, request current documentation, the DPA, subprocessor list, and relevant security materials directly.
Activepieces pricing FAQ
Is Activepieces really free?
Yes, the hosted Free plan is $0 with 100 daily credits for one user, checked 6 September 2026. Free self-hosted Community Edition provides the automation core, with infrastructure and operator costs borne by you. Neither means every paid feature is free.
Does one credit cover every step?
One credit covers a normal flow run regardless of ordinary step count. AI is additional: fast, smart, and frontier steps cost 2, 10, and 20 credits respectively under the rules checked 6 September 2026. Agentic actions cost one each.
What is the cheapest plan for five people?
Plus includes up to five users for $16/month billed yearly, checked 6 September 2026. If those five people need SSO or standard roles, Team is the first published qualifying tier at $166/month billed yearly.
Is BYOK always cheaper?
No. It lowers each AI step to one Activepieces credit on Plus and up, but adds a separate provider bill. Compare the complete monthly cost and model quality; small workloads within their allowance may not produce meaningful savings.
Sources and methodology
This is vendor-document-verified editorial analysis, not hands-on testing. Prices, credit rules, plan limits, support tiers, and edition boundaries were checked against the Activepieces pricing page and FAQ on 6 September 2026, including the full official-page content where the readable extraction omitted details. Deployment context comes from the Activepieces install overview. Internal cluster pages were checked live on the same date.
Scenario arithmetic is our own; staff rates, provider spending, infrastructure budgets, and setup hours are explicitly assumed rather than observed. The illustrations are AI-generated explanatory assets, not screenshots, benchmarks, or evidence of product testing. Recheck checkout and contractual limits before buying: pricing changes can make older indexed explanations misleading.