Workers Platform
Deploy intelligent agents that work alongside your team.
Workers are autonomous units—AI agents or human specialists—that execute tasks, monitor systems, and scale your engineering capacity. Deploy them on cloud, on-premise, or hybrid infrastructure.
Worker Templates
Pre-built workers for common automation scenarios. Deploy in minutes, customize as needed.
Adaptive Assistant
Orchestrator
Routes work, summarizes health, keeps rituals on track.
QA Sentinel
Quality
Regression sweeps and quality alerts across releases.
Code Companion
Development
PR reviews, cleanups, and architecture guide sync.
Support Champion
Support
Auto-triage, draft macros, rally specialists.
Knowledge Curator
Documentation
Curate docs, recordings, and tickets into knowledge.
Release Captain
Delivery
Launch trains, go/no-go rituals, retro scorecards.
Build Custom Workers
Don't see what you need? Build your own worker from scratch. Combine AI models, skills, triggers, and guardrails into a worker tailored to your workflow.
name: "PR Review Bot"
type: ai
model: claude-3-5-sonnet
skills:
- code_review
- security_scan
- style_check
triggers:
- event: pull_request.opened
- event: pull_request.synchronize
trust:
level: medium
require_approval:
- merge
- delete_branchHow It Works
Choose Base Type
Start from a template or blank canvas. Pick AI model or human worker.
Define Skills
Add capabilities like code generation, testing, documentation, or custom skills.
Configure Triggers
Set when the worker activates—events, schedules, webhooks, or manual.
Set Guardrails
Define boundaries, approval flows, and trust levels for sensitive actions.
Visual Builder
Drag-and-drop interface for non-technical users
YAML Config
Infrastructure-as-code for version control
Fork Templates
Start from existing templates and customize
Plugin System
Extend with custom skills and integrations
Custom AI Agents
Build AI agents with your choice of foundation model. Configure capabilities, personality, context limits, and safety boundaries.
Foundation Models
Claude 3.5 Sonnet
AnthropicBest balance of speed, cost, and capability for code tasks.
GPT-4o
OpenAIMultimodal reasoning with vision and audio support.
Claude 3 Opus
AnthropicMaximum capability for complex reasoning tasks.
Local LLM
Self-hostedRun models locally for privacy and cost control.
Bring your own API keys or use codmir-managed credits.
Agent Capabilities
Code Execution
Run code in sandboxed environments with full language support
Computer Use
Control browser and desktop for complex automation tasks
Multi-turn Conversation
Maintain context across long-running task chains
Trust & Safety
Built-in guardrails with approval flows for sensitive actions
Tool Calling
Integrate with APIs, databases, and external services
Voice Interface
Wake word activation and voice-controlled commands
Agent Personas
Pre-configured AI personalities with distinct communication styles
Efficient, precise, methodical. Code generation and task execution.
Vigilant, proactive. Monitors systems and detects issues.
Collaborative, context-aware. Interactive development assistant.
Infrastructure Options
Deploy workers where they make sense for your team. Cloud for convenience, on-premise for control, or hybrid for the best of both worlds.
Deployment Flexibility
Railway (Cloud)
Production • SaaS • Teams
Managed cloud deployment with automatic scaling, environment management, and CI/CD integration. Best for teams who want zero infrastructure overhead.
LAN Cluster
Enterprise • Security • Compliance
Distributed node agents running on your local network. Ideal for air-gapped environments, compliance requirements, or maximum control over data.
Self-Hosted (Docker)
DevOps • Hybrid • Custom infra
Deploy workers anywhere Docker runs. Each worker includes a production-ready Dockerfile. Compose files available for multi-worker setups.
Serverless (Inngest)
Automation • Events • Cost-optimized
Event-driven workers that scale to zero. Pay only when tasks execute. Perfect for bursty workloads and webhook-triggered automation.
Worker Preferences
Fine-tune every aspect of worker behavior. From triggers and permissions to notifications and resource limits—full control over your automation.
Triggers & Schedules
Define when workers activate
Trust & Permissions
Control what workers can do
Notifications
Get alerts on worker activity
Context & Memory
Configure knowledge access
Team Access
Manage who can configure workers
Secrets & Keys
Securely store API credentials
Environment Presets
Development
Verbose logs, manual triggers, sandbox mode
--env=devStaging
Production-like with approval gates
--env=stagingProduction
Full automation, strict permissions
--env=prodCustom
Your own environment configuration
--env=customReady to Deploy Workers?
Start with our templates or build custom workers from scratch. The platform handles orchestration, monitoring, and scaling.