This article covers four capacity-limited parts of the Appian Platform whose size depends on the tier: Appian RPA (robotic process automation) with its bot limits, Process HQ (process mining and reporting), Appian Portals (public web apps) and the Data Fabric row limits per record type. All four appear on the tier comparison of Appian’s pricing page, which states one pricing basis for the platform, “Priced per user, per month, per app.”[1] The public pages give the limits but not the prices or the contractual definitions that surround them.
Editions
The pricing page limits these capabilities as follows.[1]
| Capability | Standard | Advanced | Premium |
|---|---|---|---|
| RPA | 5 bots | 25 bots | Unlimited bots |
| Portals | One portal per app | Unlimited portals | Unlimited portals |
| Data Fabric | Single data source, 4M rows per record type | Multiple data sources, 20M rows per record type | Multiple data sources, no row limit per record type |
| Process HQ | Not listed | Included | Included |
| Case Management Studio | Not listed | Included | Included |
| Offline experiences | Not listed | Included | Included |
Footnote 3 of the page qualifies the RPA and portal rows: the figures represent “the entitlement per application for user licenses” at each tier, and for platform user licences “the total entitlement across all applications”.[1] A customer with several applications under application user licences can therefore hold a larger total than the single-application figure suggests, and a customer with a platform user licence holds the stated figure once.
Metrics
Bot (RPA)
The pricing unit for RPA is the bot. Appian’s documentation describes the corresponding technical unit, the robot, as an entry in the Operations Console that is tied to an agent installed on a host machine.[2] The Community Edition guide gives a plainer definition: a robot is a PC or virtual machine where robotic tasks are executed, and one can configure one robot but create an unlimited number of robotic tasks.[3] The row is Bot (RPA).
Three documented rules shape how bots are counted:
- Each agent executable file is associated with a specific robot entry in the Operations Console and can be used on only one host machine.[2]
- When an agent is downloaded, any previous installation of the same agent is invalidated, and only one agent should be installed and run per host machine and per user session.[2]
- On a Windows Server with multiple concurrent sessions, one robot can run in each session, but never more than one in the same session.[2]
Taken together, a bot count reconciles with the robot entries in the Operations Console and the host machines they run on. The pages reviewed do not say whether a bot that is registered but idle counts against the limit, or how a robot is counted when several concurrent Windows sessions run robots on one server; those points belong to the order form. The relevant rows are RPA bots are limited by tier and Each RPA agent executable serves one robot and one host machine.
Portal and rows per record type
The Portal is counted per app at Standard and is unlimited above it.[1] Rows per record type is a limit on each synced record type: 4M at Standard, 20M at Advanced and none at Premium.[1] Footnote 1 says that, for cloud customers, the single data source available at Standard is the Appian Cloud database.[1] The two higher tiers also list document record types, dynamic scaling for queries, and, at Premium only, transparent data encryption for synced record types.[1]
Counting and floors
Portals and who needs a licence
A portal is a public web app that people can use without signing in to Appian. The documentation says portals allow an organization to “Expose interfaces publicly” and share specific interfaces or forms publicly without requiring an Appian licence for every user.[7] The counting consequence is that public visitors are outside the user count, while the people who build and administer the portal remain inside it. The Standard limit of one portal per app is a limit on the number of portal objects, not on visitors. See Public portal visitors do not need an Appian licence.
Process HQ
Process HQ is the business-user workspace for process analysis and reporting. The documentation states that the capabilities it describes “are included in Appian’s advanced and premium capability tiers” and that usage limits may apply.[5] It has two functions: analyzing and optimizing business processes with low-code process mining, and building custom reports and dashboards on the data fabric.[5] It is built on the data fabric and synced record types.[5] The row limit per record type is therefore likely to be relevant to what Process HQ can analyze, although the documentation does not state that link.
Two details matter for capacity planning. Queries that Process HQ generates for no-code reporting are limited to no more than 50% of the available system resources at any time, and process analysis workloads run on the data fabric alongside the customer’s other application workloads.[5] Access is controlled by the Process HQ User role, and the Data Governor role manages which record types are available in process insights and data fabric insights.[6] The documentation lists Process HQ as available for Appian Cloud, Appian Government Cloud and self-managed environments,[5] while footnote 2 of the pricing page says some features may be unavailable or limited in self-managed deployments.[1] The documents do not reconcile these statements, so the order form and the version-specific documentation should be checked.
Data Fabric limits
Row limits apply per record type and rise with the tier. They are limits on synced record types in Appian, which cache data from source systems, and not on the size of the source systems. The practical counting question is how many rows each synced record type holds, a figure the public pages do not tell customers how to report. The Community Edition uses the same concept with a much lower figure: 25,000 rows per synced record.[3]
Virtualization and partitioning
The RPA agent rules above are the only host-level rules in the pages reviewed: one agent per robot entry, one host machine per agent, and one running robot per Windows session.[2] Java 17 must be installed on the host machine to launch the agent.[2] The pages do not define how virtual machines or Windows Server session hosts are counted against a bot limit.
Cloud and self-managed
RPA is hosted in Appian Cloud or in a self-managed environment, with local agents on the customer’s machines. The pricing page states that, starting with release 24.2, RPA is available for self-managed customers using Appian on Kubernetes.[1] In Appian Cloud HA environments, Appian RPA and trainable classification and extraction skills are not highly available, although they can still be used.[4] Autoscale for Process, listed at Advanced and Premium, needs additional licensing and, according to Appian’s Kubernetes FAQ, is not currently available for Appian on Kubernetes.[1][8] See Autoscale for Process requires additional licensing.
Programs
No programme is specific to these capabilities. The Community Edition includes one RPA robot, 25,000 rows per synced record and a daily limit on robotic task executions;[3] see Appian Trial and Community Edition terms.
Out of scope
This article does not cover AI features that rely on documents processed in Process HQ or RPA, which are metered in AI Actions (see Appian AI Actions and AI licensing). It does not cover Case Management Studio beyond its tier placement, Appian Government Cloud terms, Appian solutions sold as packaged applications, or partner and marketplace plug-ins, which the pricing page describes as not supported elements of the software.[1] Prices, the meaning of “app” for counting, and the treatment of idle or retired bots are not stated in the public documents.