Every client's requests in one place, answered in your own words.
Odly brings each client's email, Telegram, Slack and WhatsApp (on your own Meta Cloud API account) into one hub. It routes every request to the right team and drafts replies from the threads you have already resolved, separately for each client. Run it yourself or let us manage it, with your own AI model or ours. Start free.
Encrypted in transit · credentials encrypted at rest · 2FA · Self-hostable · Bring your own AI model · Bring your own storage · Bring your own database
Measured on a live client deployment (two workspaces, five people), 20 Aug – 4 Sep 2026.
Connects with the tools you already use
Odly in 60 seconds
See how Odly works
Five short drawn explainers: what Odly does, who's in control, what it costs, where your data lives and how it's protected.
The solution
What changes when requests land in Odly
Scattered channels become one system your whole business works from.
Why Odly
Built differently from the support bots
Most tools are a support desk with an AI bot added on and billed by the answer. Odly is one engine you own, pointed at every team and every client.
Each client is an isolated workspace — its own teams, and AI trained on their resolved threads.
Answers you can actually trust
Platform
What's in the platform
From the first message to resolution: routing, spam defense, knowledge and tickets, in one place and under your control.
#A1B2← they replied2h
#C3D4new12m
#E5F6→ awaiting them1d#G7H8resolved3dThe learning engine
It learns from your agents' corrections
It starts from real work: the docs, tickets and resolved threads your best agents handled. Odly drafts in their voice, reuses answers that already worked, and turns every manual correction into a rule it applies automatically.
Channels in
Email
Telegram
Slack
Chat widgetThe Odly hub
Resolved
How it learns
You choose how freely it acts
Client transparency
Your clients follow their request without logging in
When Odly auto-replies to an email, it attaches a secure tracking link. The customer opens a live status page and can reply from it, no account required.
Expected reply within an hour · Handled by the Billing team
- Received2d ago
- Reviewed & categorized2d ago
- In progressnow
- Resolved
Add to this request
Integrations
Channels and integrations
Email, Telegram, WhatsApp and Slack are all two-way: agents reply from Odly and the answer goes back where the customer wrote. Email connects over IMAP on every plan (direct Gmail sign-in on Enterprise); WhatsApp runs on your own Meta Cloud API account, with approved templates once the 24-hour window closes. An embedded widget answers website visitors, and custom API lookups pull the customer's records from your own systems. A public API is on the roadmap.
Our commitment
Your customers' messages are yours. Odly encrypts your credentials, runs on your own infrastructure if you want it, and your data is never sold or used to train anyone else's models.
Security & control
Security and data control
Built for EU agencies handling their clients' customer data. Encrypted, never sold and self-hostable, so it stays in your jurisdiction and under your control.
Pricing
One workspace with its teams
Plans are messages, with AI included on every paid plan and no per-resolution fees. Each plan has an agent allowance, from Free forever (AI on us to start: about 100 messages plus up to €1 for import and knowledge-base processing, then your own key; our database) to fully self-hosted.
A workspace is one business, client or brand, with its teams. Priced flat, never per seat. Your keys or ours (BYO AI & S3), cloud or self-hosted.
All prices exclude VAT. VAT is added at checkout where it applies. The rate depends on your country, and VAT-registered businesses elsewhere in the EU are reverse-charged.
Free, Starter, Pro and Enterprise Cloud are self-serve, so you can start in minutes. Only self-hosted and the organisation plan are sales-assisted.
FAQ
Questions, answered
What happens when Odly gets something wrong?
You decide how much it can do on its own. In Manual mode Odly only suggests; nothing sends without an agent's approval. As you build trust, proven rules can act automatically within safe limits, and you can dial that autonomy up or down at any time.
Can we switch AI-written replies off?
Yes, per workspace. Switch AI drafts off and Odly works as smart search over your knowledge base — no model writes to your customers. Auto-replies, suggested drafts and the chat widget's AI answers stop; agents still see past replies and knowledge-base matches, and the widget becomes a contact form. Routing, classification and spam checks keep using the model, so this is not "no AI at all".
Where is my data hosted?
Your choice. Run Odly as a managed service, or self-host it entirely on your own infrastructure. Either way, stored credentials and tokens are encrypted with AES-256, and your data is never sold or used to train anyone else's models.
Where are my files and attachments stored?
In storage you control. Point Odly at any S3-compatible bucket (MinIO, AWS, or your provider of choice), so attachments and knowledge-base files live on infrastructure you own, never ours. On managed plans, storage runs on Odly-managed infrastructure by default.
How long does setup take?
Every cloud plan is self-serve: connect a channel and go live in minutes on a 14-day free trial, no card. Enterprise Cloud adds hands-on onboarding on top, a session within 5 business days of signup, where we import your knowledge base and tune routing with you so Odly answers in your voice from day one. Self-hosted is the one plan we set up with you, because it runs on your own infrastructure.
Can I use my own AI model?
Yes. Bring your own: OpenAI, Anthropic, DeepSeek, Perplexity, Qwen, Bedrock, a custom endpoint, or local models like Ollama. Self-hosted deployments can keep every inference inside your own environment.
Which channels are supported?
Email, Telegram, WhatsApp and Slack are two-way: agents reply from Odly and the answer goes back into the same thread or chat. Email connects over IMAP on every plan (any mailbox, including Gmail and Google Workspace), and direct Gmail sign-in is available on Enterprise. WhatsApp runs on your own Meta Cloud API account, with approved templates after the 24-hour window. An embedded widget answers website visitors. A public API is on the roadmap.
How does pricing work?
It's a flat fee per workspace (one business, client or brand, with its teams), with no per-seat fees inside each plan's agent allowance (2 on Free, 5 on Starter, 20 on Pro, 50 on Enterprise Cloud), and never per resolution. Free is €0 forever, with AI on us to start (about 100 messages, plus up to €1 for import and knowledge-base processing) and your own AI key after that, on our database within Free's limits (or your own Postgres). Paid plans are Starter at €150/mo, Pro at €500/mo and Enterprise Cloud at €1,000/mo, all fully managed with AI included. Self-hosted is €3,000/mo on your own infrastructure. Every paid plan starts with a 14-day free trial: the full experience with up to 1,000 messages, no card.
What counts as a message, and what happens at the limit?
A message is one inbound customer message Odly analyses: an email, a Telegram or Slack message, a chat-widget request. Your agents' replies, duplicates and mail Odly filters out do not count. The allowance runs per billing period: buy on the 5th and it resets on the 5th of the next month. At the limit your inbox keeps receiving and you keep replying; only the managed AI (routing, drafts, auto-replies) pauses until the period resets, you upgrade, or you add a 1,000-message pack for €50. AI is included for every message; importing history and building the knowledge base use the same monthly allowance.
Book a demo
Book a demo
Every cloud plan is self-serve, so you can start in minutes without a demo. If you want a guided tour, we'll walk through your workflow and show how Odly would handle it.