[CNCF LFX Proposal] Meshery Grounding AI in the Meshery Registry

### CNCF Project

Meshery

### Term

2026 Term 3 (Sep-Nov)

### Program Name

Grounding AI in the Meshery Registry

### Program Description

## Description

The Meshery Registry holds thousands of components across hundreds of Models spanning Kubernetes, the CNCF landscape, and the major clouds, together with the Relationship definitions that describe how those components legitimately connect. That registry is precisely the knowledge a language model needs in order to generate a design Meshery will accept, and it is orders of magnitude too large to fit in a context window. Today there is no defined mechanism for selecting registry context for AI prompts, and even ordinary component search lacks a model filter. Generation therefore falls back on whatever the model happens to remember about Kubernetes (or any of the public Clouds that Meshery supports), which is how agents invent components that do not exist and propose relationships the policy engine will reject.

This project builds the retrieval layer that grounds every other AI capability in Meshery. The mentee implements an embedding index over Models, components, and Relationships with a pluggable backend so that operators can choose a hosted embedding service or a fully local one, incremental re-indexing as models are registered or updated, and a ranked context assembly API that accepts an intent and returns a compact, token-budgeted slice of the registry for the AI Adapter and the Meshery MCP Server to consume. The same index pays off directly for humans: semantic search across the registry pages in Meshery UI, and search improvements in `mesheryctl` including the missing model filter for component search. Effectiveness is measured, not asserted, using the design validation and evaluation work running in parallel this term.

## Expected Outcomes

- An embedding index over Meshery Models, components, and Relationships, incrementally rebuilt on registry changes, with a pluggable hosted or local embedding backend.
- A context selection API that returns a ranked, token-budgeted registry slice for a given intent, consumed by both the AI Adapter and the Meshery MCP Server.
- Semantic search in the Meshery UI registry experience and in `mesheryctl`, including the currently missing `--model` filter on component search.
- Measured improvement in component resolution rate and generated design validity against a retrieval-free baseline, reported through the term's evaluation harness.
- Documentation of the retrieval architecture, index schema, backend configuration, and tuning guidance.

### Technologies

Golang, React, vector search, REST, SQL

### Skills same as Technologies?

- [ ] Yes, the required skills are the same as the technologies listed above.

### Required/Desirable Skills

Kubernetes CRDs

### Mentors

Mia Grenell | @miacycle | mia.grenell2337@gmail.com | mia.grenell
Lee Calcote | @leecalcote | leecalcote@gmail.com | leecalcote

### Upstream Issue URL

https://github.com/meshery/meshery/issues/20995

### Application Prerequisites

- [ ] Resume
- [ ] Cover Letter
- [ ] School Enrollment Verification
- [ ] Participation Permission from school or employer
- [ ] Coding Challenge
- [x] Custom Prerequisite (fill in details below)

### Coding Challenge URL

_No response_

### Custom Prerequisite Name

Participation is Key

### Custom Prerequisite Description

1. Join the community (https://slack.meshery.io).
2. Join the weekly newcomers meeting (https://meet.meshery.io) and learn about the project.
3. Get a feel for Meshery's capabilities in the Playground (https://play.meshery.io)
1. Create a design.
2. Receive a badge (from https://meshery.io/community/recognition/).
Optionally:
1. Add badge to your GitHub profile.
2. Publish your design the Catalog.
4. Ascertain whether your passions and skills align.

### Custom Prerequisite - File Upload

- [ ] Yes - completion of this task requires the mentee to submit a file.

---
**LFX program:** [CNCF - Meshery: Grounding AI in the Meshery Registry (2026 Term 3)](https://mentorship.lfx.linuxfoundation.org/project/45b2ab48-4925-41eb-b91d-87f2e74be320)

Start with the upstream Meshery issue 20995 and review how the registry, AI Adapter, Meshery MCP Server, Meshery UI, and mesheryctl currently consume registry data. Map the retrieval, indexing, search, and evaluation boundaries before proposing an implementation. Done means incremental grounded retrieval, semantic search, the component model filter, measured evaluation results, and architecture and configuration documentation.

Written by the indexing model from the issue text.