Pinned Tweet
Agentican is now in public preview, open to everyone.
No invitation. No waitlist. No credit card. Sign up and delegate your first task to AI agents in seconds.
100+ agents across 8 departments. You describe the task. They plan and run it.
app.agentican.ai/
We're open-sourcing a multi-agent framework for Java.
Declarative plans. Parallel steps. Human checkpoints. Crash recovery. Two-line hello world.
Built from everything we learned building Agentican.
agentican.ai/blog/java-multi…
Agentican retweeted
Added support for remote MCP servers as agent tools in @agentican_ai the other day. It turns out it wasn't that hard. Maybe I'll add A2A support next.
Agentican retweeted
HITL is always after the fact. Agents act. Human provide feedback. What if agents should ask for help or additional clarification before acting? With @agentican_ai, they can... and do.
Agentican retweeted
I added support for remote MCP servers as agents tool in Agentican. Next, I'm thinking about A2A support so you can add remote agents to your team.
#aiagents
We shipped something no other agent platform has: agents that ask you questions.
Not approval gates. Actual questions — when instructions are unclear, when context is missing, when there are multiple valid paths.
Other frameworks agents guess. Ours ask.
agentican.ai/blog/human-agen…
The era of agent engineering is over.
You can't code agents to handle every task correctly — same way you can't script every decision an employee makes.
It's a management problem. Agents are a reflection of their managers.
We're all agent managers now.
agentican.ai/blog/agent-mana…
1
12
Agent frameworks want you to believe complex workflows need code.
They don't. LLMs can plan.
Describe what you need. The agent builds the workflow — loops, branches, parallel steps, approvals. Then the platform runs it.
No LangGraph. No CrewAI. No code.
agentican.ai/blog/no-code-ag…
1
23
The AI industry is obsessed with automating humans out of the loop.
Wrong goal.
The future isn't automated. It's hybrid. Humans lead. Agents scale.
We built Agentican — 100+ agents across 8 teams you manage like a workforce, not configure like software.
agentican.ai/blog/introducin…
1
2
21

