Skip to content
OpenSourceSoftware.ai
Sim logo

Sim

simstudioai · Agents

GitHub stars

29,940
—Awaiting first sync

Awaiting first sync

No measured star history yet

What is Sim?

Sim lets a team build AI agents by connecting blocks on a canvas instead of writing a full application. An agent can use company documents, structured tables and connected business apps, then be shared as a chat, an API or a tool for other AI clients. Each run is logged, so you can see what every step did and what it cost.

Sim is a workspace for building, deploying and monitoring AI agents and workflows. Workflows are built on a canvas from blocks such as Agent, Function, Condition, Router and Loop, and they connect to a catalogue of integrations and model providers. The same workspace holds tables, files and knowledge bases that agents can read and update.

Core capabilities

  • Builds workflows from Agent, Function, Condition, Router, Loop and Parallel blocks.
  • Deploys a workflow as an API, a chat interface or an MCP server.
  • Starts runs from schedules, webhooks, RSS feeds, tables or integration events.

A canvas of blocks

Workflows are built by connecting blocks on a canvas and testing each step before deploying. The Agent block calls a model and can use tools, the Function block runs your own code, and Condition, Router, Loop and Parallel blocks control the flow. Response, Guardrails and Wait blocks shape how a run behaves and ends. Integrations, external MCP servers and custom tools can be added as blocks or as tools for an agent.

  • Workflow canvas
  • Agent block
  • Integrations
  • MCP

Knowledge, tables and deployment

Knowledge bases hold searchable documents that agents use to inform their answers, tables hold structured rows that workflows can read and update, and files can be shared with agents. Triggers include Start, Schedule, Webhook, RSS, Table, Sim Workspace Events and integration triggers. Deploying creates a versioned snapshot that serves API, chat, MCP and automatic trigger runs. Logs record each block's inputs, outputs, errors, timing and costs.

  • Knowledge bases
  • Tables
  • Triggers
  • Logs

Run it yourself under Apache-2.0

The npx sim-setup wizard creates a deployment folder, generates secrets and starts the published images with Docker Compose, and a Helm chart is available for Kubernetes. The docs give 12 GiB of memory as a starting point for an evaluation install. Local models can run through Ollama or vLLM, and the built-in Chat assistant uses a Sim-managed service with its own API key, which can be turned off. The core is released under the Apache-2.0 licence, while enterprise features such as SSO, SCIM, audit logs and access control sit under a separate licence that needs a paid subscription for production use.

  • Self-hosted
  • Docker
  • Kubernetes
  • Apache-2.0

Where it fits

Use cases

  1. 01

    Answer questions from company documents

    Upload documents to a knowledge base so agents can search them and retrieve the relevant passages. Add an Agent block with access to that knowledge base, test it with the Run button, then deploy the workflow as a chat interface for colleagues. Logs record each block's inputs, outputs, errors, timing and costs.

  2. 02

    Run a scheduled task across business apps

    Start the workflow with a Schedule trigger, then add integration blocks for services such as Slack, Notion, HubSpot or Salesforce. Use a Condition or Router block to choose a branch and an Agent block to summarise or sort the data. Deploy the workflow so the schedule runs the deployed version, and redeploy after each change.

  3. 03

    Let other systems and AI tools call a workflow

    Deploy a workflow as an API that other systems call, or as an MCP server that an MCP client can use as a tool. Deploying creates a versioned snapshot, so edits to the draft do not affect the live version until you deploy again. A Webhook trigger lets outside systems start the same workflow.