- Home
- Trust & Security
Trust & Security
Enterprise controls, engineered by finance people.
CoaleTech builds on the Frappe Framework and ERPNext, a platform designed for governance, with role-based access, immutable change history, and API-first integration built in. We add finance-grade audit trails and a governed-AI layer on top, so your systems are controlled, auditable, and safe to put in front of a regulator.
18 production applications · open-source core · v16 target, v15 migration.
Deployment
Your data, where you want it
Frappe is open source and self-hostable. There is no per-seat license lock-in and no obligation to run on someone else's cloud. You choose the deployment model that fits your sovereignty and cost requirements.
On-premise
Run the entire stack inside your own data centre for full data sovereignty. Nothing leaves your network.
Private cloud
Deploy into your own cloud tenancy. You own the infrastructure and the data; we engineer and operate the application.
CoaleTech-hosted
Managed hosting on our infrastructure, with scheduled backups, security patching, and monitoring handled for you.
Platform controls
Governance built into the foundation
The controls below are not add-ons. They are properties of the platform every CoaleTech application inherits.
Data isolation & infrastructure
Every deployment runs on Frappe Bench with multi-tenant sites and a Docker-ready footprint. Each site has its own isolated MariaDB database; Redis backs the cache and job queue.
Access control & authentication
Enterprise RBAC out of the box: Frappe roles, role and user permissions, and field-level permissions restrict who sees and edits what. Two-factor authentication is supported for privileged access.
Audit trails & versioning
Built-in document versioning records every change to every record. On top of that, our apps write app-level immutable action logs, such as the Agent Action Log, that cannot be edited after the fact.
APIs & integrations
Every DocType is a REST resource, with whitelisted RPC for controlled server methods. Access is authenticated with API tokens or OAuth. Integrations with banks, KRA, M-Pesa, WhatsApp, and telematics are first-class, not bolted on.
Scheduler & background work
Scheduler events and RQ background jobs run recurring and long-running work off the request path (telematics sync, tax submission, reminders) so the user-facing system stays responsive.
Backup & disaster recovery
Bench-level scheduled backups with client-configurable retention and off-site storage. Recovery objectives and failover are documented and drilled so business continuity is assured, not assumed.
Governed AI
AI you can put in front of an auditor
Numbers are computed deterministically in SQL and Python; a large language model only narrates them. It never invents a figure. Every automated action passes through the governance gate, and every invocation is written to an immutable audit log.
Green. Auto. Low-risk actions execute automatically within policy.
Amber. Confirm. Actions that change data or money require a human to confirm before they run.
Red. Review-only. High-risk actions are surfaced for review; the agent never executes them on its own.
- AAL-0041close-briefGREENauto-run · logged
- AAL-0042variance-checkAMBERconfirmed · logged
- AAL-0043je-draftREDheld for review
Engineering discipline
Built to be maintained, not just shipped
- Extension over core modification: ERPNext is extended via custom fields and doc-event hooks, never patched in place, so upgrades stay clean.
- Migration discipline: versioned patches and dedicated migrations for safe, reversible upgrades from v15 to v16.
- Code-quality tooling: pre-commit with ruff, eslint, prettier, and pyupgrade enforced across the codebase.
Due diligence
Documents we can share
Request any of these at sales@coale.tech with the deployment model you are evaluating.
We do not claim certifications we do not hold; we show you the controls instead.
Reviewing us for a security or audit assessment?
We'll walk your team through deployment options, access controls, audit trails, and our governed-AI architecture in detail.