Agent Deployment

Deploy your AI agents to production, manage versions, roll back changes when needed, and publish across multiple environments and channels.

app.8bit-ai.com
Agent deployment

Deploying an Agent

Once your agent is configured and tested, deployment makes it available to handle real conversations. The deployment process is instantaneous — your agent goes live within seconds of clicking the deploy button.

Prerequisites

  • Agent must have a name and system prompt configured
  • At least one LLM provider must be active
  • Knowledge bases (if connected) must be processed and ready
  • Voice pipeline (if enabled) must be fully configured

Deployment Methods

Dashboard Deployment

Navigate to the agent detail page and click the Deploy button. The agent status changes from "Draft" to "Active" immediately.

Status Flow: Draft → Deploying → Active (success) or Failed (error)

API Deployment

Deploy programmatically as part of your CI/CD pipeline.

Deployment Environments

8bit-ai supports multiple deployment environments for staged rollouts:

EnvironmentPurposeTrafficLimits
DevelopmentActive development and testingInternal onlyNo rate limits
StagingPre-production validationTest usersStandard limits
ProductionLive user trafficAll usersPlan-based limits

Environment Isolation

Each environment has isolated configuration, knowledge bases, and channel connections. What you deploy to staging does not affect production until you explicitly promote it.

Channel Deployment

Deploy your agent to one or more channels simultaneously. Each channel can have its own configuration overrides.

Version Management

Every deployment creates a new version of your agent configuration. Version management lets you track changes, compare configurations, and maintain a complete audit trail.

Version Tags

Each version is automatically tagged with a semantic version number and can include custom labels for easier identification.

Version History

Access the complete version history from the agent detail page. View what changed between versions, who made the changes, and when deployments occurred. The dashboard highlights differences in system prompts, model selection, and parameters.

Comparing Versions

Use the API to programmatically compare two versions and identify differences.

Automatic Versioning

Every deploy or configuration change creates an automatic version snapshot. You never lose the ability to see exactly what configuration was active at any point in time.

Rollback

If a deployment introduces issues, you can instantly roll back to any previous version. Rollbacks are immediate and restore the exact configuration from that version.

Dashboard Rollback

Navigate to the Versions tab, find the version you want to restore, and click Rollback. Confirm the action and the agent is restored immediately.

API Rollback

Roll back programmatically for automated incident response.

Rollback Considerations

  • Rolling back restores the configuration, but active sessions continue with the current version until they end
  • Channel configurations (web widget, Telegram, etc.) are also reverted
  • Knowledge base links are restored to the state they were in at that version
  • All rollbacks are recorded in the audit log for compliance

Environment Management

Manage multiple deployment environments for staged rollouts and testing.

Promoting Between Environments

After validating in staging, promote a version to production with a single action. This deploys the exact same configuration that was tested.

Environment Variables

Each environment supports custom environment variables that override agent settings for that specific deployment context.

Promotion Approval

Enterprise plans support required approval workflows for production deployments. Configure approval gates in Organization Settings to prevent unauthorized deployments.

Deployment Status Monitoring

Monitor the health and status of your deployed agents in real-time.