Content backend with typed lifecycle enforcement. Your AI agents operate on the same backend as your team, with the same rules.
Self-hosted. Open source. Zero runtime dependencies. MCP-native.
#ai #selfhosting #eutech #golang #opensource #smeldr
| web | https://smeldr.dev |
| X | https://x.com/getsmeldr |
In Smeldr, an AgentJob has the same lifecycle as any content type: Draft, Published, Archived.
Published means active. Archived means stopped. To pause an agent, archive its job. To resume, publish it again.
No code change. No redeploy. No configuration to update.
https://smeldr.dev/solved/pause-one-agent?utm_source=mastodon&utm_campaign=pause-one-agent
Publishing a batch of posts? Archive the social agent before you start. Only that step pauses.
The other agents were never connected to it. They each subscribe to their own content type and their own signal. One goes quiet. The rest keep running.
Bring it back when you are ready.
https://smeldr.dev/solved/pause-one-agent?utm_source=mastodon&utm_campaign=pause-one-agent
Smeldr gives your AI the same content model as everything else. One backend, four audiences: browser, API, AI agent, AI reader.
Same lifecycle. Same role-based access. An agent at author level can create and publish. It cannot delete. It cannot touch content it was not given access to.
One backend. One lifecycle. One set of roles.
Browser, API, CLI, and AI agent operate on the same content model, with the same rules. Your team and your AI work from the same place. No exceptions for anyone.
Self-hosted. Open source. Zero runtime dependencies.
https://smeldr.dev/solved/should-i-use-smeldr?utm_source=mastodon&utm_campaign=should-i-use-smeldr
Your AI agent can write a post. The question is what happens to that output.
In most setups, it lands in a document or a file. Someone picks it up and decides what to do with it. The agent never knows.
Smeldr closes that loop. Agent output goes through the same lifecycle as anything a human creates.
Smeldr is a content backend with a native AI interface. In fact, every interface.
Typed lifecycle enforcement, MCP-native tooling, and a backend your AI agents can operate on directly. The same rules apply to everyone, including the AI. Self-hosted. Open source. Zero runtime dependencies.
This is what forge-cms.dev became.
smeldr.dev?utm_source=mastodon&utm_campaign=this-is-smeldr
