Every plan has an annual platform fee and a monthly fee for each agent running in production. Connectors, users and sandbox agents are unlimited, and you bring your own LLM.
A fixed-fee pilot in your own systems. Pick one queue, such as late purchase orders or won-deal hand-offs, and see real agent runs from week two.
For mid-sized companies on Acumatica, QuickBooks, HubSpot or Salesforce running a few focused agents.
Platform fee + per production agent
For companies on Salesforce or SAP running agents in finance, sales operations and supply chain.
Platform fee + per production agent
Everything in Growth, plus:
For enterprises running many agents across business units, regions or several ERPs.
Platform fee + volume pricing per agent
Everything in Business, plus:
You pay for the platform and for the agents doing real work. Everything you need to build and test them is included.
One annual fee per company. It covers the agent core, every connector, approvals, the audit trail and the governance controls.
A monthly fee for each agent running live, such as one that chases unconfirmed purchase orders. Build and test as many as you like in the sandbox at no charge.
Connect your preferred model provider with your own account and keys. We recommend the models that work best for each job, and we never resell or mark up model usage.
The things that usually cost extra are included from the first day.
SAP, Salesforce, HubSpot, Acumatica, QuickBooks, Microsoft 365, any REST or OData API and any MCP server. We never charge per connector.
Everyone who briefs, approves or reviews an agent gets access, with no seat fees.
As a fully managed service, in your own cloud tenancy or on-premises, with your keys and the data residency you choose.
Approvals, segregation of duties, scoped service accounts, rate limits and a one-click kill switch.
Failed runs and retries never count toward an agent's usage.
| Feature | Growth | Business | Enterprise |
|---|---|---|---|
| Platform | |||
| Unlimited connectors | |||
| Unlimited users and sandbox agents | |||
| Bring your own LLM | |||
| Approvals, audit trail and kill switch | |||
| Build, version and promote your own agents | – | ||
| Security and deployment | |||
| Fully managed service, your own cloud or on-premises | |||
| Single sign-on | – | ||
| Separate test and live environments | – | ||
| Support and commercial | |||
| Standard support | |||
| Uptime SLA | – | – | |
| Named customer success manager | – | – | |
| Volume pricing for agents | – | – | |
| Multi-year price lock | – | – | |
Book a short call. Tell us which systems you run and which queues you want agents to take on, and we send a quote for the platform fee and the agents you need.
One job brief running live with its own service account, for example "chase unconfirmed purchase orders" or "turn won deals into ERP orders". Agents you build and test in the sandbox are free.
No. You connect your own model provider and pay them directly. Fidevra does not resell or mark up model usage. We recommend the models that work best for each type of job, so you can balance quality and cost.
Each production agent covers a generous monthly volume of completed actions. An agent that runs well beyond that counts as an extra agent. Failed runs and retries never count.
If you continue after the pilot, the full pilot fee is credited against your first year.
Yes. Our engineers add AI to legacy systems and build custom AI solutions on the same platform. Custom work is scoped and quoted per project. See services.
Yes. Move up a plan whenever you need single sign-on, separate environments, or your own agents. Your agents carry over.