Features

Every feature, by edition.

Build an AI workflow once. Deliver it to your clients as a secure app that returns the results.

Build an AI tool three ways

Feature SMB Enterprise
Prompt (natural language) — Describe the outcome in plain English; the AI pulls from your connected systems, works out the steps, and returns the result. The fastest way to ship a report. Runs read-only.
LSPL (LogicScript Prompt Language) — Write the process as a structured script — explicit steps, branches, validation — that the AI executes in order, reasoning at each step. Repeatable and unambiguous: it reads the same to any two people. Runs read-only.
Script (Python) — Full, deterministic control for exact calculations, data reshaping, and logic the AI shouldn't improvise. Runs in a secure sandbox, can call the model where you want judgement. Runs read-only.

Platform features

Feature SMB Enterprise
Self-service inputs
Pre-run input forms — Add a short form (date ranges, accounts, free text) that end users fill before clicking Run.
Typed, validated fields — Text, number, yes/no, date, dropdown, and user-picker fields are validated before the run starts.
Dynamic options — Dropdowns populate from a connected system, keeping forms accurate without manual upkeep.
Connect to your systems of record
NetSuite — Connect your ERP for revenue, financial, and operational reporting.
Salesforce — Pull CRM data — pipeline, accounts, opportunities — into governed reports.
Maxio — Subscription billing and revenue summaries from your Maxio account.
QuickBooks Online — Connect each client's QBO company for accounting reports and reconciliations.
Xero — Connect each tenant's Xero organization for accounting data.
One-click OAuth — Most connectors authorize through a secure provider login, with no API keys to copy.
Per-connection access control — Each connection is read-only by default.
Depth over breadth — Deep back-office finance and operations integrations rather than shallow connections to hundreds of apps.
Results, execution & delivery
Multiple output formats — Return Excel, formatted HTML, print-ready PDF, charts, plain text, or JSON.
In-browser results — Results render with searchable tables and clickable links back to the source system.
Asynchronous run queue — Every run processes in the background, so long reports never tie up a browser tab.
Complete run history — Every run is recorded with its inputs, result, and cost to re-open, re-run, or audit.
Scheduled runs — Run a tool automatically on a daily, weekly, or monthly cadence in your timezone.
Secure result sharing — Share a report via a secure link that opens an access-controlled viewer, never a raw file.
Email notifications — Optional per-workspace emails notify users and admins when a run finishes or fails.
AI economics & control
Per-tenant usage & cost tracking — Every run records token usage and estimated cost, rolled up per workspace.
Model tiers — Route simple jobs to a fast model and hard jobs to a frontier model, per tool.
Batch processing — Run eligible jobs in batch mode at roughly half the cost when speed isn't critical.
Bring your own key (BYOK) — Run all AI on your own provider key, keeping spend and data flow under your control.
Per-run budgets & caps — Fail-closed ceilings on tokens and calls per run stop a runaway job from running up a bill.
Model allow-listing — Only approved models can be used.
Governance, security & trust
Role-based access control (RBAC) — Grant precise per-tool permissions for who can see versus run each tool.
Single Sign-On (SSO) — Your team signs in with the credentials they already use.
Complete audit trail — Every run, connector call, and admin change is recorded in an audit log.
Always read-only — All tools are read-only and cannot update your core systems.
Secure, sandboxed execution — Python tools run in an isolated sandbox with no access to platform secrets.
Multi-tenant isolation — Every workspace's data is isolated at the database level, eliminating cross-tenant leakage.
Encryption at rest — Connector credentials and AI keys are encrypted at rest and never returned to the browser.
Multi-tenancy & the channel
Workspace isolation — Each client is a fully isolated workspace with its own data, connectors, tools, users, and results.
Client Manager console — Manage all client workspaces, staff, and invitations from a single console.
Build once, deliver to many — Author a tool once and distribute it across client workspaces.
Import / export tools — Move a library of AI Tools between workspaces as portable files.
Per-client controls — Each client carries its own settings, branding, AI key, notification policy, and grants.
Staff & invitations — Invite staff and client users by email, assign roles, and require approval where needed.
Cost attribution — Per-tenant usage and cost tracking makes spend transparent per client.
Administration
Workspace settings — Company profile, notification preferences, AI key, and capability settings per workspace.
User management — Invite teammates, approve membership, assign roles, and disable accounts.
Roles — Define roles and grant each the precise set of tools and permissions it should have.

Ship your first one-click tool today.

Stand up a secure, multi-tenant portal for your prompts, scripts, and agents in under an hour.