What's new at Cyndra

Workflows: build multi-step automations your AI employee runs on its own. Start from a gallery of ready-made templates for common jobs, or describe your own in plain language.

Triggers: start a workflow on a schedule, from an incoming app event, or with a secure webhook, so recurring work happens without anyone asking.

Approvals: require a human yes or no before sensitive steps run, with a dedicated approval queue and a full run history you can retry.

Custom tools: connect your own internal tools and services so agents can act in the systems you already use, beyond the 1,000+ built-in integrations.

Roles and audit: create custom team roles and review a searchable audit log of every action taken across your workspace.

Shareable conversations: send a read-only link to any chat so teammates can follow along.

Image attachments: drag and drop screenshots or images straight into chat and your agent can see and work with them.

Sidebar signals: see at a glance which chats are still working and which need your review, and pin the ones you use most.

Built for mobile: the full dashboard and chat now work cleanly on phones, including edge-to-edge iPhone screens.

Queue and steer: send follow-up messages while your agent is working and it handles them in order, or redirect it mid-task.

Watch it work: live progress streams into the chat, with a collapsible drawer of the steps your agent took.

Agent avatars, smart reply suggestions, and cleaner formatting for code, images, and long answers.

One-click copy, collapsible code blocks, click-to-zoom images, and a jump-to-latest button.

Live activity in chat channels: watch the steps your agent takes in Telegram, Slack, WhatsApp, and Discord, collapsed into a tidy summary when it finishes.

Video understanding: send a video, or several at once, and your agent can watch it and work with what it sees.

Cleaner, quieter status updates in customer chats, in plain language with no technical noise.

Acts only when you ask: asking about an earlier image or video no longer triggers a new one, and the agent decides when to use a tool instead of reacting to certain keywords.

Stays on task: fixed cases where an agent could stop partway through or repeat itself.

Grounded answers: connected-data reports are built from data your agent actually pulled, not guesses.

Slack and Microsoft Teams agents now respond to anyone in your workspace by default.

Per-agent host isolation: every agent now runs under its own isolated host account in a rootless container, part of our SOC 2 program.

Privacy hardening: closed edge cases that could reveal sensitive details in chat, and tightened multi-factor settings for white-label clients.

Operator console: redesigned with audit logging and clearer access controls for chat channels.

Your first AI employee is included with every plan, so you can put an agent to work from day one.

Connect-a-tool bonus: link your first app and get credits to try real work right away.

Smoother onboarding and checkout: set up billing during onboarding, with clearer steps and a prompt to enable each newly connected app on your agents.

A friendlier welcome: a guided email sequence with a link to book an onboarding call.

Connect WhatsApp with reliable pairing, now shown right on the agent page alongside Slack, Microsoft Teams, Telegram, and Discord.

Skills that grow: your agent can learn and reuse new skills on its own.

Fresh reads every time: agents pull current data from your connected apps on each request, across every channel.

Your clients sign in on your own domain, fully under your brand, with no Cyndra branding anywhere.

Guided custom-domain setup with automatic DNS, so going live on your domain is quick and clear.

Stronger brand and workspace isolation so every client stays fully separated from the others.

Clearer partner admin with plain-English buttons and inline help throughout.

Pay as you grow: no platform fees until you launch your first client AI employee.

Usage auto top-up so your client agents never pause for credits.

Redesigned launch checklist and first-run experience, so getting a client live is calm and obvious.

Simpler, transparent white-label pricing.

Files deliver as real attachments: documents, images, and other files your agent creates now arrive as proper attachments in chat.

Live step-by-step progress on chat channels while your agent works on something longer.

One clear, plain-English progress update instead of technical logs and duplicate status messages.

Send a photo and your agent understands it: improved image understanding across every channel.

Core productivity skills are now enabled by default on every agent.

Evidence-gated review: sensitive actions can be held for approval with a clear record of what the agent did and why.

More reliable media handling and a wider set of built-in skills.

AI employees that live where your team already works: Slack, Microsoft Teams, WhatsApp, Telegram, and Discord.

1,000+ integrations through one-click OAuth, plus 113 built-in skills shipped with every agent.

White-label and reseller program: run AI employees for your clients under your own brand.

Three approval modes (Supervised, Semi-autonomous, Autonomous), isolated cloud containers per agent, encryption in transit and at rest, and full audit logs of every action.

Server-side credential handling: OAuth tokens and API keys are encrypted at rest and never exposed to the agent. Cyndra does not use your data to train shared models without your explicit opt-in.