Feature Request: Official OpenViking integration for agent context and memory

[githb-ac]

Before filing

• I searched open and closed issues for duplicates.
• I am using a recent release and this doesn't already exist.
• This is one request, not several bundled together.

Closest existing issue

none found

Is this new, or an improvement?

New capability — Berd can't do this at all today

The problem, in your terms

Berd agents currently do not have a built-in way to use OpenViking as a shared context, memory, knowledge, and skills backend.

OpenViking is designed to organise agent memory, knowledge, and skills with hierarchical retrieval. Without an official Berd integration, users must configure OpenViking separately, create their own adapter or MCP connection, manage credentials, and troubleshoot compatibility themselves.

This makes it difficult for Berd agents to reliably remember previous work, past decisions, recurring mistakes, and shared team knowledge across sessions and workspaces.

What you do today

I would have to run OpenViking separately and connect it to Berd manually through an MCP-compatible configuration or an unofficial wrapper.

That means the user must manage:

• The OpenViking deployment.
• The connection endpoint.
• Credentials.
• Agent and workspace permissions.
• Compatibility and error handling.
• Updates to both Berd and OpenViking.

There is currently no clear Berd-native setup or status indicator for this.

What you'd like to see

Add an official OpenViking integration to Berd that:

• Provides OpenViking in Berd’s context, memory, or integration settings.
• Supports both local/embedded and remote OpenViking deployments.
• Allows secure entry of the endpoint and credentials.
• Detects whether OpenViking is installed and reachable.
• Exposes OpenViking memory, knowledge, and skills to Berd agents.
• Allows users to scope OpenViking to selected agents, projects, or workspaces.
• Shows clear connection, authentication, compatibility, and error status.
• Documents supported OpenViking versions, deployment modes, and limitations.

If a full native integration is not appropriate, an official and documented MCP-compatible adapter for Berd would also solve the problem.

Why this belongs in Berd itself

Berd is intentionally extensible, so this could potentially be implemented as a skill, agent, extension, or automation.

However, an official Berd integration would still be valuable because it could provide:

• A discoverable setup inside Berd.
• Secure credential handling.
• Agent and workspace scoping.
• Connection and health checks.
• Consistent error messages.
• Version and compatibility checks.
• A supported integration path that does not depend on unofficial wrappers.

A user can create a basic external connection today, but that does not provide the integrated setup, status, permissions, and support experience described above.

Non-goals

This request is not asking Berd to:

• Replace its existing agent or context system.
• Force every Berd user to install OpenViking.
• Replace Berd’s skills or extensions.
• Replace Buzz’s separate OpenViking integration request.
• Provide a hosted OpenViking service.
• Store OpenViking data on Berd’s servers.
• Remove the option to use other memory or context providers.

The request is for a user-controlled integration.

Alternatives you considered

Possible, but it requires users to manage endpoints, credentials, scope, and errors manually.

Mockups, prior art, or other context

No response