Pricing
Four self-serve tiers (Dev $99 to Scale $2,999) plus custom Enterprise — usage-metered on compute units, with a 14-day free trial on the Dev plan. Annual prepay saves 10% — 18% and 25% on 2- and 3-year terms. BYOK on every tier — bring your own Anthropic, OpenAI, or Bedrock-EU keys and pay your provider directly.
Usage-based pricing
Pay for what your assistant does — never a surprise bill.
Syncanix meters usage in compute units (CU) — one weighted number for the real work your assistant does: messages, tool and MCP calls, and the tokens underneath. You pay for usage, not per seat, and every action's weight is published up front.
- 2 CUChat message answered
- 2 CUTool or MCP call
- 1 CUEndpoint action run
- 5 CUGenerated insight or report
Plus the model tokens underneath. On Managed plans, the model tokens each answer uses meter at about 1 CU per $0.0016 of model cost — a short answer on the default model adds roughly 1 CU, while a long answer that works through a lot of context on a top-end model can add 30 or more. On BYOK, model tokens draw 0 CU and bill straight to your own provider.
What a month can look like
- A quick question your assistant answers on its own is about 3 CU — a Dev plan's 9,000 CU covers a few thousand of those a month.
- A typical support chat — the agent answers (2 CU), calls one tool to look up the visitor's order (2 CU), plus the tokens underneath — is around 6 CU. A busy month of 4,000 of those is about 24,000 CU, comfortably inside Starter's 30,000.
- A long, multi-step answer that reads a lot of context and calls several tools on a top-end model can be 30 to 80 CU — the same allowance covers far fewer of those, which is exactly what the live usage gauge is for.
- Every weight is published and tracked on a live usage gauge in your dashboard — no black box.
- At your included CU the assistant pauses by default, so a runaway loop can never run up a surprise bill.
- Want to keep going? Opt in to pay-as-you-go overage at $0.012 per unit — billed next invoice, never retroactively.
Managed by default: we hold the model keys and token cost is bundled into your CU. Prefer your own? Bring your Anthropic, OpenAI, or Bedrock-EU keys on any tier and your tokens bill straight to your provider.
Starter
$299/mo30,000 CU
- 5 seats
- BYOK or managed LLM — model usage included
- email support
Growth
Most popular$999/mo110,000 CU
- Frontier included — Copilot Mode, Change Preview + Undo, automations
- 10 seats · all 4 chat surfaces
- MCP server · drift detection
- BYOK — bring your own LLM keys
- multilingual agent — 6 languages, RTL included
- priority support · branding removed
Scale
$2,999/mo360,000 CU
- Frontier included — Copilot Mode + automations
- 25 seats
- BYOK recommended — managed available
- custom theming · 4h-response SLA (business hours, CET)
Enterprise
Customcustom
- VPC · SSO/SAML · DPA
- SOC 2 Type I (Q3 2026) · custom SLA · CSM
- self-hosted MCP · multi-year rate locks
Just starting out? Dev — $99/mo · 9,000 CU · 2 seats · 14-day free trial, no card
talk to sales- 14-day trial — no card to start
- cancel anytime — your catalog stays exportable
- hard-stop billing by default — no surprise overages
No free tier — every plan starts on Dev with a 14-day trial, no card required.
Every plan's included compute units are published — nothing is unlimited, no bill cliffs. Run out and you can opt in to pay-as-you-go overage at $0.012 per unit, or upgrade for a bigger bucket.
CU — compute units. You pay for what your assistant does (messages, tool calls, tokens), not per user.
Estimate your monthly cost
Drag the slider to your expected monthly interactions — a message and the tool calls and tokens it drives — and we'll estimate your compute units and recommend a plan.
Estimated compute units: 30,000 CU / mo
About 6 compute units per interaction at this weight — actual usage depends on the messages, tool calls, and tokens each one drives.
Recommended plan
Starter$299/mo
or $269/mo billed annually — save 10%
30,000 compute units included every month
Past your included allowance, usage stops by default — no surprise bill. Opt in to pay-as-you-go overage at $0.012 per unit to keep running, billed on your next invoice.
Estimates only, for planning. Your live usage gauge and the published weight table show exactly how each message and action draws compute units.
Compare every plan
Every capability, side by side. The numbers come straight from the plan configuration — what you see here is what each plan enforces.
| Feature | Dev $99 | Starter $299 | Growth $999 | Scale $2,999 | Enterprise Custom |
|---|---|---|---|---|---|
| Usage & limits | |||||
| Included compute unitsper month | 9,000 | 30,000 | 110,000 | 360,000 | Custom |
| ≈ Interactions per monthapproximate, at a typical interaction weight | 1,500 | 5,000 | 18,333 | 60,000 | Custom |
| Team seats | 2 | 5 | 10 | 25 | Custom |
| Pay-as-you-go overageopt-in; charged only past your included allowance | $0.012/CU | $0.012/CU | $0.012/CU | $0.012/CU | $0.008/CU |
| Product & surfaces | |||||
| Embeddable chat widget | Included | Included | Included | Included | Included |
| Per-tenant MCP server | Included | Included | Included | Included | Included |
| Multilingual — 6 languages, RTL | Included | Included | Included | Included | Included |
| End-user authentication (act-as-user) | Included | Included | Included | Included | Included |
| Action confirmations + step-up | Included | Included | Included | Included | Included |
| Server-side executionrun MCP tools against your own API (Channel B) | Not included | Not included | Included | Included | Included |
| Remove “Powered by Syncanix” branding | Not included | Not included | Included | Included | Included |
| Frontier — Copilot ModeChange Preview + Undo, generative dashboards, automations | Not included | Not included | Included | Included | Included |
| Data, security & trust | |||||
| EU data residency (Frankfurt) | Included | Included | Included | Included | Included |
| Tenant isolation (Postgres RLS) | Included | Included | Included | Included | Included |
| Conversation data retentiondays; Enterprise default 90, configurable by contract | 7 | 30 | 90 | 365 | 90 |
| Support | |||||
| pricing.featureMatrix.row.support.label | Community | Priority | 4-hour SLA (CET business hours) | CSM + custom SLA | |
Uniform capabilities are available on every plan — no plan gates the widget, MCP server, multilingual support, end-user authentication, action confirmations, EU data residency, or tenant isolation.
Frequently asked questions
- What is a compute unit (CU)?
- A compute unit is a weighted measure of what your assistant actually does — each message, each tool or MCP call, and the underlying tokens all consume CU. You pay for usage, not per user: a power user and a casual user are billed by what they actually do, never a flat per-head fee.
- What happens when I run out of compute units mid-month?
- No surprises — we never bill retroactively. You get dashboard and email warnings at 50%, 80%, and 95% of your included compute units. At 100%, your assistant pauses by default and your data and conversation history stay readable — or you can opt in to pay-as-you-go overage and keep running, billed at $0.012 per extra unit on your next invoice. No surprise invoices, ever.
- Does BYOK actually save me money?
- Your plan price doesn't change — BYOK changes what's on the usage meter, not your subscription. On Managed plans we hold the model keys and fold token cost into your compute units, so it's one predictable invoice. With BYOK you bring your own Anthropic, OpenAI, or Bedrock-EU keys: your tokens bill straight to your provider at your negotiated rate and draw 0 CU, so your monthly allowance stretches to cover more real usage. BYOK is recommended at Scale and Enterprise, and managed access stays available on every tier.
- What is the annual prepay discount?
- Annual prepay saves 10% across Starter, Growth, and Scale; 2- and 3-year prepay terms save 18% and 25%. Enterprise rates are negotiated; multi-year locks follow the same schedule.
- How do I cancel?
- Self-serve in your dashboard at any time. Monthly plans stop at the end of the current billing period. Annual plans are non-refundable but you keep service through the paid term. Your catalog stays exportable forever; no lock-in.
- What is included with Enterprise?
- VPC deployment, SSO/SAML, signed DPA, the SOC 2 evidence pack (Type I targeted Q3 2026), a custom signed SLA, a dedicated customer success manager, self-hosted MCP option, and multi-year rate locks. Custom pricing on a committed compute-unit contract — contact sales for a tailored quote.
- Does a heavier user cost more than a light one?
- Yes — that's the point. Because you pay by compute units, a user who sends 100 messages and triggers dozens of tool calls draws far more CU than one who asks a single question, and your bill reflects that. The live usage gauge and the published weight table show exactly how each message and action consumes CU, so heavy usage is transparent, never a surprise.