Sim

Sim is an open-source workspace for building, running and operating AI agents. It collapses the usual split between a visual automation builder and a codebase: a chat panel, a node graph and the execution logs all describe the same workflow, so you can describe an automation in plain English, watch it materialise as connected blocks, then open any block and edit the prompt, model or code behind it.

What you build with it

A workflow starts from a trigger — an inbound webhook, a schedule, a chat message, a form — and chains together model calls, API calls, conditionals and loops. Agent blocks can be pointed at any major LLM provider, and Sim ships more than 1,000 integrations so a single run can enrich a lead from a CRM, summarise it with a model, write the result to a table and post a notification to Slack without leaving the canvas.

The workspace around the canvas

Sim is more than the editor. A knowledge base holds documents that agents retrieve against, Tables give workflows structured state that persists between runs, Files handles uploads, and Scheduled Tasks runs workflows on a cadence. Logs record every execution — inputs, outputs, latency and cost per block — which is what makes a workflow debuggable once it is running unattended rather than a black box.

Deployment and licensing

The project is open source on GitHub and can be self-hosted, or used as a managed cloud workspace. The free tier includes 1,000 monthly credits, 10 concurrent executions and 5 GB of storage; Pro is $25 per user per month and Max is $100, with an Enterprise tier for custom credit pools, higher concurrency and unlimited log retention.

Who it suits

Teams that have outgrown no-code automation but do not want to maintain agent orchestration infrastructure — the visual layer keeps the workflow legible to non-engineers while the code layer keeps it precise enough for production.