Kaer Is Now in Public Beta: Why We're Building in the Open
Today, Kaer is open to everyone. No waitlist, no invite codes. You can sign up, create your first agent, and start running tasks in about two minutes.
We could write one of those polished launch posts — the kind that makes it sound like everything was planned from the start and every feature is perfect. But that would be dishonest, and honesty is something we take seriously.
So instead, here's the real state of things.
What we've built
Kaer is a platform for deploying autonomous AI agents. You describe a task in plain language, choose an execution model, and let the agent handle it. Agents can search the web, execute code, read and write files, make HTTP requests, scrape websites, and send emails. They plan their approach, use tools, handle errors, and deliver results.
Beyond single tasks, we've built three additional capabilities:
Automations — Schedule agents to run on cron schedules, in response to webhooks, or continuously. A competitor report every Monday morning. A daily digest of relevant industry news. An automated response to inbound leads.
Workflows — Chain multiple agents together with conditional logic, branching, and parallel execution. One agent's output feeds into the next. This is where simple tasks become powerful processes.
Assistants — Deploy always-on AI bots to messaging platforms. Telegram, Discord, WhatsApp, or custom webhook endpoints. They carry persistent context and can use the same tools as regular agents.
What's working
Agent quality is where we've invested the most, and it shows. Our agents reliably handle multi-step tasks that would take a human 30-60 minutes. Research tasks — "analyze these three companies" or "find recent papers on X" — are particularly strong. Code generation and execution work well for scripting, data analysis, and automation tasks.
Reliability has been solid. We're at 99.7% uptime since launch, and first-token latency is consistently under two seconds. When an agent hits a dead end — a search returns nothing useful, an API is down — it adapts and tries alternative approaches rather than failing silently.
What needs work
We're being honest here because we'd rather you know what to expect.
Long-running tasks (over 10 minutes) occasionally lose coherence. The agent does the work, but the output quality degrades as the task complexity increases. We're actively working on better context management and sub-agent decomposition to fix this.
The workflow builder is functional but young. Some of the more advanced routing patterns need polish. We ship improvements to it weekly.
Our connector ecosystem is limited. Right now, connecting to third-party tools requires webhooks or API calls. We're building native integrations with the platforms people use most — starting with Slack, GitHub, and Google Workspace.
Why we build in the open
Every company says they value transparency. We try to actually practice it. Our roadmap is public. Our changelog is updated with every release. When something breaks, we say so.
This isn't altruism — it's strategy. The AI space is full of products that overpromise and underdeliver. We'd rather set accurate expectations and earn trust through consistency. If an agent can't do something well, we'd rather tell you upfront than have you discover it during a critical workflow.
Where we're headed
Over the next three months, we're focused on three priorities:
- Agent quality. Improving long-running task performance, adding better reasoning capabilities, and expanding the tool ecosystem.
- Integrations. Native connectors for the 20 most-requested platforms, starting with Slack, GitHub, Jira, Linear, and Google Workspace.
- Team features. Shared workspaces, role-based access, audit logs, and SSO. The features that let companies adopt Kaer across their entire organization.
If any of this interests you, the best way to stay updated is to sign up and try it. The Starter plan is free and includes enough credits to run real tasks — not just a demo.
We're building something we think matters. We're glad you're here for it.