AGEFreighter for Visual Studio Code

Plan, run, recover, and review validated graph migrations without moving migration logic or credentials into an AI model.

AGEFreighter migrates CSV, PostgreSQL, Neo4j, and Azure Cosmos DB graph data to Apache AGE or PostgreSQL 19 SQL/PGQ property graphs. This extension is a guided interface for the separately installed AGEFreighter 2.3.0 CLI; the Go CLI remains the deterministic migration engine and owner of durable checkpoints.

Highlights

Discover AGEFreighter LoadJob YAML and JSON files in the workspace.
Validate configuration and inspect static plans without connecting to a source or target.
Run bounded source profiles and target readiness diagnostics.
Start and resume migrations in a visible terminal.
Review durable status, verification, migration reports, and optimization recommendations.
Ask @agefreighter to explain bounded evidence using the chat model selected in VS Code.
Expose one confirmed, read-only tool to VS Code agent mode. AI tools never start, resume, clean up, or mutate a migration.

Prerequisites

Install AGEFreighter from the Visual Studio Marketplace, then install the AGEFreighter 2.3.0 CLI and ensure agefreighter is on PATH. Alternatively, run AGEFreighter: Select CLI Binary and select the executable.

On macOS, the Homebrew installation is:

brew install rioriost/cask/agefreighter

Linux and Windows release archives are available from the AGEFreighter releases. Windows CLI binaries in 2.3.0 are provided without an Authenticode signature; verify their checksum and GitHub build-provenance attestation before use.

Start a migration

Open a trusted workspace containing an AGEFreighter LoadJob file.
Open the AGEFreighter activity-bar view.
Expand a discovered job and run Validate, then Static plan.
Use Profile source and Diagnose target when their connection cost is acceptable.
Select Start migration, inspect the confirmation, and continue in the visible terminal.
Keep the UUID printed by load. Status, resume, verify, report, and cleanup require that durable job ID.

Long-running commands are intentionally opened as terminal processes. Reloading the extension does not silently resume or create another migration. Use durable status and the CLI's reviewed resume procedure after a failure.

AI assistance

The extension works without GitHub Copilot or another chat model. If VS Code chat is available, use @agefreighter /help or one of the read-only slash commands. The selected model may explain evidence and recommend a next step, but all execution remains in the CLI and all mutations require a direct modal confirmation outside chat.

Only bounded, recursively redacted CLI JSON is eligible for model context. Raw job files, environment variables, connection strings, credentials, credential reference names, queries, source records, stderr, and terminal logs are never sent by this extension. See Privacy and security.

Settings

Setting	Default	Meaning
`agefreighter.binaryPath`	`agefreighter`	Executable path or command name
`agefreighter.readTimeoutSeconds`	`120`	Timeout for captured read-only commands
`agefreighter.maxOutputBytes`	`4194304`	Per-stream output capture limit

Workspace trust and remote development

Job discovery works in restricted mode. No AGEFreighter process runs until the workspace is trusted. The extension runs where the workspace extension host runs, so Remote SSH, Dev Containers, and Codespaces need the AGEFreighter CLI installed in that remote environment.

Virtual workspaces are not supported because the CLI requires filesystem paths.

Support

AGEFreighter is open source under the MIT License.