Early days, and moving fast
Seline is under heavy, active development. Things will change, occasionally break, and sometimes disappear as we iterate in the open. Expect rough edges — and expect them to get smoother quickly.
You're trying an early preview. Your feedback directly shapes where the platform goes next.
Built for real autonomy
Always Online
Stays active around the clock, continuously monitoring data and events with configurable heartbeats and scheduled automation.
Interrupts
Instantly pause or stop any ongoing task without friction, giving you full control over execution at every step.
Connectors
Seamlessly integrates with multiple data sources and AI tools, unifying context across platforms into a single workflow.
Entity Graph
Maintains a structured graph of relationships, enabling deeper context awareness and more accurate reasoning over time.
Sub Agents
Breaks complex tasks into coordinated sub-agents, providing visibility, parallel execution, and better control over progress.
Second Brain
Continuously captures, organizes, and recalls information, building a reliable memory layer that evolves with your workflows.
How it works
From sign-up to a working agent in about a minute. Here’s the whole path.
1Sign up
Create your account in seconds. No credit card, no setup wizard to fight with.
2Connect your accounts
Link a Telegram bot and add an API key for your model provider. Your keys are stored encrypted.
3Seline spawns your agent
A personal container boots up just for you — isolated, always-on, and yours alone.
4Talk to Anamika
Message your agent on Telegram like you would a colleague. Ask, delegate, automate.
What can you ask Seline?
Just ask in plain language and Anamika gets to work — searching the web, reading and summarizing pages, running quick commands, and keeping notes for you.
“Search the web for a good starter DSLR under $500 and summarize the options.”
“Read this article and pull out the three key takeaways for me.”
“Look up today’s weather in Bangalore — will I need an umbrella?”
“Save these meeting notes to a file so I can pull them up later.”
“What did I write down about the trip plan? Pull up my notes.”
“Split a ₹4,200 dinner three ways with a 10% tip — what’s each share?”
Connects to what you already use
Bring your own bot and your own model. Seline plugs into your accounts and pulls context across platforms into one workflow.
More connectors are on the way — Seline is built to grow with your stack.
Your data, your container
Seline handles your API keys and personal context, so trust is the product. We built it to be private by default — isolated, encrypted, and permissioned.
Your own container
Every user gets a fully isolated agent container. Your data and your agent never share space with anyone else’s.
Encrypted secrets
API keys and bot tokens are encrypted at rest. They live in your container and are never exposed in plain text.
You stay in control
An allowlist gates who can message your agent, and nothing destructive happens without your explicit confirmation.
Fresh from the blog
Product updates, stories, and what we’re building next.
Frequently asked questions
Everything you might want to know before spinning up your agent.
How much does Seline cost?
Seline is in early access while we finish polishing things. You bring your own model API key, so your only model cost is what your provider charges you directly. Pricing for the hosted, always-on container will be announced before we exit early access.
Can I self-host it?
Yes. The agent and server are designed to run in your own environment. If you’d rather not manage infrastructure, the hosted version spins up and maintains your personal container for you.
What does "Always Online" cost in resources?
Your agent runs in a lightweight container that idles cheaply between tasks and only does real work when something happens — a message, a scheduled job, or a heartbeat check. You control how aggressive the scheduling and monitoring are.
How is my data retained?
Your conversation history and context live inside your own isolated container. Secrets are stored encrypted. Older history is compacted over time to keep memory relevant, and you can clear it whenever you want.
Which model providers are supported?
Out of the box you can connect OpenRouter, Sarvam, NVIDIA, or Cerebras by adding an API key. OpenRouter is the default and gives you access to a wide range of models.
How do I talk to my agent?
Through Telegram. You connect a bot during onboarding, add your username to the allowlist, and then message Anamika like you would any other chat — no special app to install.