Updated 2 November 2026. Ansible is an independent messenger with its own server and clients based on open source. It runs in a preproduction environment; security and compatibility verification are ongoing.
The Erlang/OTP server uses MVSy, a protocol in the MTProto 2.0 family with a TL schema. The current server layer is 232; older-layer compatibility is checked separately. See the schema, methods and layers.
TCP, HTTP and WebSocket transports and obfuscation modes are available. Transport encryption and end-to-end encryption are different: cloud conversations sync through the server, while secret-chat keys stay on devices.
Services run in Kubernetes in one region. Control-plane high availability and backup recovery remain separate work areas; multi-region resilience is not claimed.
Website assets are delivered through a CDN. Messenger media services use separate routes and access checks; the website CDN does not establish coverage of every messenger media scenario.
my.ansible.su manages applications and API credentials. See OAuth documentation for website sign-in and the bot platform for automation.
Public links use asme.su. They open in a browser and offer a route into the Ansible client.
Diamonds are an internal currency. Balances and the financial journal reside in the server database; wallet services, API and UI use a shared payment path. Atomic operations and replay protection cover a number of transfers, purchases, upgrades and gift transfers. Remaining operations and result delivery are still being verified.
Cryptocurrency, TON and Toncoin withdrawals are not part of the product. Real payment providers are disabled. Compatible Stars/XTR API names do not imply cryptocurrency.
Secret chats use device-held keys and do not migrate history through cloud sync. The group-call SFU is deployed; call scenarios still require checks on particular clients and devices.
A schema specifies request and response formats, not full functional parity. There is no verified overall coverage percentage: checks account for layers, participant roles, permissions, retries, live updates and subsequent reads.
We do not claim a completed security audit, full compatibility on every client or a guaranteed SLA.