1. Create an agent
In Runtite, an agent is an AI model with capabilities, triggers, and instructions.
Start by selecting the model to use, which determines the available integrations.
Shared AI workplace for teams
Adopt AI agents responsibly
Run agents in a shared workplace with your team.
Chat with them, automate tasks, and review their work.
What is Runtite?
A platform where agents work with your team
Runtite is an agent workplace. Your team can create AI agents using any model provider (or your own self-hosted models) and interact with them to get work done.
It allows your team to connect integrations, manage agent permissions centrally, and keep an audit trail of all agent activity.
How it works
From zero to agent in minutes
General integrations
- API endpointIntegrate with any API.
-
GmailSearch emails, make drafts, and send mail. -
Google CalendarManage calendar events.
Provider built-in capabilities
Web search Fetch up-to-date information via the model provider's native search
Image generation Create visual assets
Code execution The model can generate code and run it in the provider's sandbox
MCP server Connect external tools through OpenAI and Anthropic MCP support
2. Connect integrations
Give capabilities to your agent. Enable any of our custom integrations, and/or the model provider's built-in capabilities.
3. Add triggers
Make your agent available for your team to chat with, configure it to run on a schedule, or get a unique URL to trigger it on-demand.
You can give it different instructions for each of these triggers.
That's it!
A new agent has joined your team. You can chat with it, and review its work when it runs in the background.
Chat with your agents
Use any AI model, commercial or self-hosted
Use the latest models from OpenAI, Anthropic, Google, and more,
or connect to your own self-hosted models with an OpenAI-compatible interface.
Streaming conversations
Persistent chat history with files, images, and source citations.
Context at a glance
See how much of the model's context window a thread is using as it grows.
Decisions in the thread
Approve or deny agent actions without leaving the chat.
Trigger automations
Automate work in the background
Configure agents to run on a schedule, or trigger them from external systems via webhooks.
The agent executes in the background and, if human approval is required, adds a task for your team to review.
On schedule
Webhook
... More triggers coming soon
Each execution is recorded in an audit trail. See who did what when, and how much it cost.
Checks and balances
Quality over quantity
Generative AI models are eager to, well, generate a lot of output. But how do you control the quality of what's generated?
Runtite gives your team the tools to do so.
Human-in-the-loop approval
Centrally configure which integrations may run automatically and which of them require approval.
When an action needs intervention, Runtite creates a task in the inbox. Your team can review it, and the agent will pick up where it left off.
Gmail (send emails)
Google Calendar (create, update, delete events)
Gmail (read emails)
Google Calendar (read events)
Runtite divides integrations in tools and data sources so you can configure permissions for read-only actions separately, even if they belong to the same integration.
Stay in control
Adopt AI in your team with guarantees that give you assurance.
Centralized credentials
Connect Gmail, Google Calendar, and other OAuth integrations once per organization.
You can centrally rotate or revoke them, and that will apply to all agents using them.
Runtite (and your agents) never see your passwords.
Usage dashboard
Track token usage, agent activity, flow volume by trigger type, pending approvals, and more.
Example breakdown
Clear pricing
You pay a flat fee to use Runtite (we don't charge per user). On top of that, you pay for your model usage.
In the future, we want to offer managed credits for your convenience. For now, you can bring your API keys.
During Runtite's beta phase, you only pay for your model usage. The Runtite platform itself is free.
Use cases
Example workflows your team could set up in minutes
Inbound lead triage and follow-up drafts
Webhook or form submissions kick off an agent; Sales reviews scored leads and drafts before the agent sends them a follow-up email.
Meeting prep from calendar context
Agents pull upcoming events from Google Calendar, scan related threads in Gmail, and draft a summary for your team before the call.
Post-meeting discussion summary
After a call, your team stores the meeting notes in your internal system. The agent then takes them, pulls context from the meeting invite, and drafts a summary that your team reviews before sending.
Account and prospect research briefs
Web search and uploaded notes become talking points your team verifies before outreach.
Weekly competitor or market pulse
Scheduled runs summarize changes on web search results; Marketing reviews the digest before sharing it internally.
Weekly team status digest
Scheduled agents send a digest to the team on what was worked on last week.
Support inquiry triage from Gmail
Agents classify incoming requests in your inbox, pull context from your APIs and web results, and propose responses. Your team reviews before the agent replies to your customers.
CRM updates through your APIs
Agents call your CRM's endpoints to log notes or update records after a customer event happens.
Triage new Sentry bugs
New Jira or Sentry bugs trigger an agent. It gathers context from the Sentry issue, Jira and your code repo, and writes a comment to the ticket with an initial assessment.
FAQ
Some more information about Runtite
Join the waitlist to get notified when we launch
Runtite is in active development. We will send you an email when we open up the private beta.