Buddy Works
The agentic cloud. Compute, storage & delivery in one place.
149 followers
The agentic cloud. Compute, storage & delivery in one place.
149 followers
Host apps and websites, spin up Sandboxes, manage domains, run tests, deploy to any cloud, and automate delivery with pipelines, all in one platform. Configure everything through the UI, YAML, or code, or let Claude Code, Cursor, Copilot, and other AI agents operate Buddy directly through MCP or CLI, including deployments, infrastructure, and logs.
This is the 5th launch from Buddy Works. View more
Buddy AI Access (MCP)
Launching today
Give AI agents secure access to your infrastructure and workflow. With Buddy MCP agents can run pipelines, inspect logs, deploy apps, create sandboxes, publish artifacts and manage domains. Control exactly what they can access with workspace permissions and tool scopes. Works with Claude Code, Cursor, Codex, Copilot and other MCP clients.
Free Options
Launch Team
Buddy Works
When an agent triggers a run or a deploy, who shows up in the audit log - the agent, or me, because it's my token?
Buddy Works
@tprus If you log in as yourself during MCP authorization, you're marked as the invoker. If you log in as an agent (bot user), the action is assigned to the agent
Buddy Works
@tprus It depends on how you authorized to MCP, as yourself or as the agent. If you authorized as yourself, you show up as the user who performed the action. If the authorization was done as the agent, then the agent is shown instead. The agent's own log entry records which user authorized as it at that time.
Buddy Works
Hi Hunters! 👋
I'm Paul, one of the makers behind Buddy.
We use Buddy to build Buddy, so AI access has been a part of our own workflow for quite some time.
One of the things I use it for the most is debugging failed CI builds. It allows me to check the logs from my terminal, figure out what went wrong, push a fix and run the pipeline again without jumping between various tools. The productivity boost is tangible.
The same goes for working with targets. Being able to safely run commands on them means debugging becomes part of the flow instead of a disruption.
We're excited to finally share this with all of you. Would love to hear how you're using MCP clients in your own workflows!
what about the CLI. many of us are using terminal agents like claude code or codex. is there a CLI for those cases?
Buddy Works
@pawel_grobelniak In a terminal agent you can use both: MCP or CLI.
The bdy CLI give you pretty much the same capabilities. From my perspective, CLI is the better fit for terminal agents
https://buddy.works/docs/cli/reference
How granular are the permissions for AI agents? Can you limit an agent to specific pipelines, environments or actions?
Buddy Works
@magdalena_anderson
Great question!
It’s very granular, and you can control access on two levels:
1. You can control which resources the agent can access. Create an Agent in Buddy, which works like a bot user, and assign it only to the projects, pipelines, and targets it needs. You manage it just like a regular workspace member. Then connect MCP as that agent rather than as yourself, otherwise it inherits everything your own account can access
2. You control what it can do. OAuth and personal access tokens let you limit scopes by resource type, such as pipelines, code, targets, or environments. For example, you can give it read only access if that’s all it needs
Toki: AI Executive Assistant
Congrats on the launch! Giving AI agents controlled access to real infrastructure feels like a big step beyond just code assistance. Curious — how granular can the permissions get when teams want to limit what an agent can deploy, inspect, or modify?
@jamieg
There are two layers of security you can introduce:
First, which resources it can access. Create an Agent in Buddy and assign it only to the projects, pipelines, and targets it needs.
Second, what it can do. OAuth and personal access tokens let you limit scopes by resource type, including read only access.