Skip to main content

Phase 1: Preparation

1

Obtain replicated credentials

  • Contact Qodo to receive your Replicated registry credentials
  • Verify access to both registries:
    • artifacts-self-hosted.qodo.ai
    • artif-reg-self-hosted.codium.ai
2

Prepare infrastructure

  • Deploy and configure Kubernetes cluster
  • Set up kubectl access with admin permissions
  • Install helm cli.
  • (For Context Engine) Deploy PostgreSQL v17+ with required databases and pgvector
3

Configure external dependencies

For all products:
  • Set up AI model access (API keys for OpenAI, Anthropic, Vertex AI, or AWS Bedrock)
  • Configure Git provider webhooks and authentication.
For Context Engine specifically:
  • Create PostgreSQL databases for indexer and metadata service
  • Configure network access from K8s cluster to database
  • Set up vector database (pgvector)
4

Gather configuration values

  • AI model API keys
  • Git provider credentials and webhook secrets
  • Database connection strings
  • Domain names and TLS certificates
  • Organization/team structure

Phase 2: Deployment

General Helm deployment pattern: All Qodo products follow a similar deployment workflow:
Product-specific deployment steps: Qodo IDE:
  • Create .secrets.toml file with AI model keys
  • (Optional) Enable Agentic Mode with additional database configuration
  • (Optional) Enable CronJobs and Jobs for Agentic Mode database partitioning
Qodo GIT:
  • Create .secrets.toml with Git provider configuration
  • Configure webhooks in your Git provider
  • Deploy with analytics sidecar
Context Engine (Multi-Service Deployment): Deploy each service separately with its own values file:
  1. Metadata Service
  2. Indexer Service (with reindex CronJob)
  3. Context Retriever Service
  4. Configure shared secrets for all services

Phase 3: Configuration

1

Secrets management

All products use TOML files for secret configuration, mounted as Kubernetes secrets:Example structure:
2

Ingress configuration

Example ingress configuration (EKS with ALB):
3

Git provider integration

For GitHub:
  • Create GitHub App with appropriate permissions
  • Generate private key
  • Install app on organization/repositories
  • Configure webhook URL to point to your ingress
For GitLab/Bitbucket:
  • Configure webhook secrets
  • Set up bearer tokens or OAuth
  • Configure API base URL for self-hosted instances

Phase 4: Validation

1

Verify pod status

2

Check service endpoints

3

Test connectivity

  • For Qodo GIT: Trigger a test webhook from Git provider
  • For Qodo Context: Add test repository to index
  • For Qodo IDE: Connect IDE and test chat/generation