# Relay blog

> Engineering notes, product updates and guides from Relay: database access for people and AI agents, with guardrails, masking and audit.

Each post has a Markdown copy at its URL plus `.md`. Product: https://userelay.sh. Docs: https://docs.userelay.sh.

## Guides

- [How to use Claude Code agents with Relay](https://blog.userelay.sh/guides/how-to-use-claude-code-agents-with-relay.md): Install the Relay skills with relay skills add --agent claude-code, open the database with relay connect, and point Claude Code at the local port it prints. Claude never holds a password. Relay checks, masks and logs each statement, and risky writes wait for a person to approve them.

- [How to use Codex agents with Relay](https://blog.userelay.sh/guides/how-to-use-codex-agents-with-relay.md): Install the Relay skills with relay skills add --agent codex, run relay connect for the database, and describe the local port in AGENTS.md. If the Codex sandbox blocks the connection, allow network access for workspace-write. Codex never sees a password, and Relay checks, masks and records every statement it sends.

- [How to use PlanetScale with Relay](https://blog.userelay.sh/guides/how-to-use-planetscale-with-relay.md): Create a branch password in PlanetScale, register it once with relay connections create, and open it with relay connect. People and agents then use a local port. Relay keeps the PlanetScale password, connects over TLS, masks sensitive columns and records every statement against the person who opened the session.

- [How to use Supabase with Relay](https://blog.userelay.sh/guides/how-to-use-supabase-with-relay.md): Create a dedicated Postgres role in Supabase, register the session pooler address as a Relay connection with Supabase's CA certificate, and open it with relay connect. Clients and agents use a local port, row-level security still applies to the role, and Relay masks and audits every statement on the way through.