Oracle AI Data Platform (AIDP) Developer Extension

Build, manage, test, and deploy Oracle AI Data Platform workflows from VS Code.

The Oracle AI Data Platform (AIDP) Developer Extension connects VS Code to Oracle AIDP so you can browse workspaces, edit files, manage agents, run Workflow jobs, inspect job runs, work with compute resources, and operate catalog resources without constantly switching to the AIDP console.

Features

• Browse workspace folders, files, agents, compute groups, Workflow jobs, Workflow job runs, knowledge bases, and master catalog resources.
• Preview remote files, download editable local copies, compare with the server version, re-download from the server, and upload changes back to AIDP.
• Create workspace folders, create agents, create .py, .sql, and .txt files, and manage supported workspace resources from the IDE.
• Create and manage agent files, set semantic roles such as entry file and dependencies file, attach or detach AI compute, deploy to Playground or PROD, and copy deployed endpoint URLs.
• Test deployed agent endpoints with built-in Debug Chat, request and response inspection, debug entries, trace visualization, and optional trace analysis.
• Execute .py and .sql files remotely on attached compute, reuse or interrupt active execution sessions, and inspect SQL results, Spark progress, and execution output.
• Run Workflow jobs, open the created Job Run in a separate tab, inspect task results, and delete obsolete Job Runs.
• Browse and manage master catalog resources, standard catalogs, schemas, managed volumes, volume files, tables, and knowledge bases.
• Notebook support is planned for a future release.

Install

Install from the VS Code extension marketplace:

1. Open Extensions in VS Code.
2. Search for Oracle AI Data Platform (AIDP) Developer Extension.
3. Select Install.

You can also install from the command line:

code --install-extension oracle.aidp-code-editor

Prerequisites

• VS Code 1.122 or later.
• An OCI API signing key registered with your Oracle Cloud account.
• Access to an Oracle AI Data Platform data lake and workspace.
• Permission to browse and manage the AIDP resources used by your workflow.

Setup Wizard

The setup wizard guides you through authentication, AIDP punchout URL, and workspace selection. You can get started by clicking the Begin Setup button in the Setup & Status section.

Agent and compute selection are handled later from the Workspace tree, Setup & Status actions, and deployment flows.

Browser Login

Browser Login is the recommended flow. It opens your browser, completes sign-in, and returns the token to VS Code. Tokens are stored in VS Code SecretStorage, while shared non-secret Browser Login profile data is stored under $HOME/.aidp-vscode/aidp-code-editor.config.json.

Browser Login setup flow:

1. Pick Browser Login.
2. Provide the AIDP punchout URL.
3. Select the workspace.

API Key Flow

The API Key flow signs requests with an OCI config profile. Use this flow when your environment requires API-key request signing.

API Key setup flow:

1. Pick API Key.
2. Set up OCI credentials.
3. Provide the AIDP punchout URL.
4. Resolve the aidataplatform alias.
5. Select the workspace.

OCI Credentials

The extension uses $HOME/.oci/config by default, or the path from aidp.ociConfigPath. The config file must point to your OCI API key pair.

Generate an API key if needed:

1. Log in to the OCI Console.
2. Click your profile icon, then open My profile.
3. Scroll to API keys and click Add API key.
4. Choose Generate API key pair or upload your own public key.
5. Download the private key and save the .pem file somewhere safe.
6. Click Add.
7. Copy the generated configuration file preview.

Create $HOME/.oci/config with the copied configuration:

[DEFAULT]
user=ocid1.user.oc1..aaaaaaaxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
fingerprint=ab:cd:ef:12:34:56:78:90:ab:cd:ef:12:34:56:78:90
tenancy=ocid1.tenancy.oc1..aaaaaaaxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
region=us-chicago-1
key_file=$HOME/.oci/oci_api_key.pem

Copy the private .pem file into $HOME/.oci/ and make sure key_file points to it. Use an absolute key_file path to avoid ambiguity when opening different workspaces.

IDE Navigation Panels

The AIDP activity bar container provides the main extension views:

• Workspace: Browse folders, files, agents, compute groups, Workflow jobs, and Job Runs directly from the IDE sidebar.
• Master Catalog: Explore catalogs, schemas, tables, volumes, and volume files without leaving the editor.
• Setup & Status: View setup status, active workspace, selected agent or Workflow item, available compute, deployment endpoints, Debug Chat actions, and selected Job or Job Run status.

Browse Your Workspace

Explore AIDP workspace resources directly in the VS Code sidebar. The Workspace tree supports browsing folders, files, agents, compute groups, Workflow Jobs, and Job Runs without switching to the AIDP console.

Local File Mapping And Synchronization

• Preview remote AIDP files in VS Code.
• Edit local copies of AIDP files.
• Re-download AIDP files from the server, overwriting the local copy.
• Compare local files with the latest AIDP version.
• Upload one or more local files, overwriting AIDP versions.
• Reveal local copies in Finder or the file explorer.

Workspace Authoring

• Create workspace folders.
• Create agents.
• Create .py, .sql, and .txt files.
• Notebook support is planned for a future release.

Workspace Utilities

• Refresh the workspace.
• Open related AIDP pages using Manage on Server for supported Workspace, Compute, Workflow, and Master Catalog nodes.
• View default agent status in the VS Code status bar and Setup & Status section.

Agent

The Agents section supports browsing, managing, and deploying agents in VS Code.

Create And Delete Agents

• Create agents from the Workspace tree.
• Optionally select AI compute during agent creation.
• View indicators for agents with or without attached compute.
• Open or view agents using Manage on Server.
• Delete agents with confirmation.

Agent File Management

• Upload local files into an agent.
• Create, rename, and delete files inside an agent folder.
• Preview, download, compare, re-download, and upload agent files.

Agent Semantic Configuration

• Set a .py file as the Entry File.
• Set requirements.txt as the dependency file.
• View visual indicators for semantic roles.

Deployment Lifecycle Management

• Deploy to Playground and PROD.
• Redeploy deployments.
• Undeploy from Playground, PROD, or all environments.
• Test and validate deployed agents.
• Poll for deployment status.
• Copy deployed endpoint URLs.
• Upload files before deploy.
• Select AI compute during deployment.
• View attached compute state in the Workspace tree and Setup & Status section.

Compute Attachment

• Attach or detach AI compute from an agent.
• View Clusters and AI Computes under Workspace -> Compute.
• Inspect compute details.
• Start, stop, or restart eligible compute resources.
• Open Compute pages using Manage on Server.

Debug Chat And Endpoint Testing

The Debug Chat section supports testing deployed agent endpoints, inspecting request and trace details, and using AI-assisted trace analysis in VS Code.

Chat

• Start endpoint test chat from a deployed endpoint.
• Open UI chat mode for agents.
• Select the active Playground or PROD endpoint.
• Start a new chat thread.
• Send chat messages and follow-up actions.
• Configure debug chat settings, including model override and extra JSON context.

Request, Response, And Trace Inspection

• View the chat transcript for endpoint interactions.
• Inspect raw request and response payloads.
• View debug entries and log output. The Debug tab shows messages emitted with the __aidp_debug__ prefix.
• Visualize execution traces.
• Track streaming or incremental trace updates.
• Copy deployed endpoint URLs.

AI-Assisted Trace Analysis

• Analyze traces with OCI Generative AI.
• Configure the trace analysis model.
• Control trace analysis settings and policy through extension configuration.

File Execution

Remote file execution runs supported files on AIDP compute rather than on the local machine.

• Attach a cluster.
• Execute .py and .sql files remotely.
• Reuse, interrupt, or close active execution sessions.
• View SQL results, Spark progress, and execution output in integrated views.
• Notebook support is planned for a future release.

Jobs

The Workflow Jobs section supports browsing jobs from the Workflow listing API, previewing .job files, running jobs, tracking execution status, and viewing job run results.

• Browse Workflow Jobs with a configurable list limit.
• Preview Job details, including tasks, compute, dependencies, and file locations.
• Run a Workflow Job from the tree or Setup & Status banner.
• Open the newly created Job Run in a separate tab when it becomes available.
• Browse Workflow Job Runs.
• Open Job Run result details.
• Delete Job Runs.
• Open Workflow pages using Manage on Server.

Master Catalog

The Master Catalog section supports browsing catalogs, schemas, tables, and managed volumes, along with catalog administration and volume file management in VS Code.

Catalog

• Create Standard Catalog.
• Create Schema.
• Browse schemas and tables.
• Open Master Catalog pages using Manage on Server.

Managed Volume Operations

• Create and delete volumes.
• Browse volume folders and files.
• Preview volume files and open related items in the browser or AIDP console.
• Upload, download, delete, re-download, compare, and push volume files.
• Reveal local copies of volume files in Finder or the file explorer.

Knowledge Base

The Knowledge Base section supports browsing knowledge base sources, previewing source files, uploading knowledge base documents, and running ingest operations in VS Code.

Common Actions

Settings

Plugin Architecture And Concepts

These are the core interaction models behind the extension.

• Local Mapping Model: Remote workspace and volume files can be opened as local editable copies in VS Code. The extension maintains a mapping between local files and their remote AIDP paths so changes can be compared, uploaded, re-downloaded, and tracked through local synchronization workflows.
• Remote Execution Model: Supported file execution runs on AIDP compute rather than on the local machine. The extension uses cluster-backed sessions and session reuse to execute .py and .sql files remotely while returning results to VS Code. Notebook support is planned for a future release.
• Agent Deployment Model: Agent deployment is managed through structured lifecycle actions in the extension. Users can validate flows, select Playground or PROD deployment, attach compute when needed, and monitor deployment progress and endpoints directly from VS Code.
• Knowledge Base Workflow: Knowledge base operations are organized around source-backed ingestion workflows. Users browse source volumes and files, upload documents, and trigger ingest operations that move content into the knowledge base indexing lifecycle.
• Debug And Trace Workflow: Debugging is centered on integrated endpoint testing. The extension sends requests through embedded Debug Chat, exposes raw payloads and trace details, and supports AI-assisted trace analysis to help users inspect runtime behavior.
• Shared Configuration Model: Shared Browser Login flow state and plugin-level profile data are stored under $HOME/.aidp-vscode, while workspace-local config is reserved for data that belongs to the opened folder.

Target Users

• AIDP developers building and deploying agents.
• Technical product teams validating agent behavior.
• Data and AI engineers working with AIDP workspaces, Workflow Jobs, Job Runs, compute resources, and storage volumes.
• Internal teams using AI coding assistants to automate AIDP development workflows.

Troubleshooting

If setup fails or the extension cannot connect, open View -> Output and select Oracle AI Data Platform (AIDP) Developer Extension.

Feedback And Issues

This extension is in preview. If you run into problems or have feature requests, contact the extension support team.