weave-setup

name	weave-setup
description	Set up and configure weave-cli for vector database RAG solutions. Use when: (1) installing or updating weave-cli, (2) creating or updating config.yaml and .env files, (3) running weave doctor diagnostics, (4) selecting and configuring vector databases, (5) verifying health of VDB connections. NOT for: data ingestion (use weave-ingest), querying (use weave-search), evaluations (use weave-eval), or stack deployment (use weave-stack).
homepage	https://github.com/Maximilien-ai/weave-cli
tags	["rag","data","weave","vector-database"]
metadata	{"openclaw":{"emoji":"⚙️","requires":{"bins":["weave"]},"install":[{"id":"brew","kind":"brew","formula":"Maximilien-ai/tap/weave-cli","bins":["weave"],"label":"Install weave-cli (brew)"},{"id":"go","kind":"go","package":"github.com/Maximilien-ai/weave-cli","bins":["weave"],"label":"Install weave-cli (go install)"}]}}

weave-setup

Set up and configure weave-cli for building multimodal RAG solutions across 10 vector databases.

When to Use

• Installing weave-cli for the first time
• Creating or updating config.yaml and .env for a project
• Running diagnostics to verify everything works
• Selecting which VDB(s) to use based on project requirements
• Checking VDB connectivity and health

When NOT to Use

• Data ingestion, collections, documents -> weave-ingest skill
• Searching, querying, agents -> weave-search skill
• Evaluations and benchmarks -> weave-eval skill
• Stack deployment (k8s/podman) -> weave-stack skill
• End-to-end planning -> weave-planner skill

Installation

# Clone and build from source
git clone https://github.com/Maximilien-ai/weave-cli.git
cd weave-cli && ./build.sh
# Binary at: bin/weave

# Verify
weave --version
weave doctor

Configuration

Interactive Setup

weave config create --env       # Create config.yaml + .env interactively
weave config update --env       # Update existing configuration

Manual config.yaml

quiet-config: true
logging:
  level: info

databases:
  default: weaviate-cloud
  vector_databases:
    - name: weaviate-cloud
      type: weaviate-cloud
      url: ${WEAVIATE_URL}
      api_key: ${WEAVIATE_API_KEY}
      openai_api_key: ${OPENAI_API_KEY}
      timeout: 10
      collections:
        - name: MyDocs
          type: text
          description: Main document collection

.env File

OPENAI_API_KEY=sk-proj-...
WEAVIATE_URL=https://your-cluster.weaviate.cloud
WEAVIATE_API_KEY=your-key
# Add credentials for whichever VDBs you configure

Diagnostics

weave doctor                     # Full system diagnostic
weave doctor --fix               # With auto-fix suggestions
weave doctor --section config    # Check specific section
weave doctor --section vdb       # VDB connectivity only
weave doctor --verbose           # Show passing checks too
weave doctor --json              # Machine-readable output

Doctor sections: system, config, env, vdb, llm, embeddings, stack, opik

Health & VDB Info

weave health check               # Quick VDB health
weave health check --weaviate-cloud
weave vdb list                   # List all configured databases
weave vdb info weaviate-cloud    # Details on specific VDB

Config Management

weave config show                # Display current config
weave config list                # List configured databases
weave config sync                # Sync to ~/.weave-cli
weave config fix --errors-only   # Auto-fix issues
weave config show-schema         # Show config schema

Supported Vector Databases (10)

VDB	Local	Cloud	Best For
Weaviate	weaviate-local	weaviate-cloud	Production, all features
Qdrant	qdrant-local	qdrant-cloud	Rust performance, filtering
Milvus	milvus-local	milvus-cloud	Horizontal scaling
Chroma	chroma-local	chroma-cloud	Simple setup (macOS)
Supabase	supabase-local	supabase-cloud	PostgreSQL users
Neo4j	neo4j-local	neo4j-cloud	Graph + vector hybrid
MongoDB	mongodb-local	mongodb-cloud	Atlas Vector Search
Pinecone	—	pinecone	Auto-scaling managed
Elasticsearch	elasticsearch-local	elasticsearch-cloud	Hybrid search
OpenSearch	opensearch-local	opensearch-cloud	AWS, k-NN + BM25

VDB Selection Guide

• Quick prototyping: Weaviate Cloud (free tier) or Chroma local
• Production: Weaviate, Qdrant, or Milvus
• Existing infra: Match your current DB
• Graph + vector: Neo4j
• AWS native: OpenSearch

VDB Override Flags

weave --weaviate-cloud <command>
weave --milvus-local <command>
weave --qdrant-cloud <command>
weave --all <command>          # All configured VDBs
weave --mock <command>         # Mock for testing

Global Flags

--config FILE       # Config path (default: ./config.yaml)
--env FILE          # Env path (default: ./.env)
-v, --verbose       # Verbose output
-q, --quiet         # Minimal messages
--json              # JSON output
--no-confirm        # Skip confirmations
--timeout DURATION  # e.g., 5s, 30s

Decision Workflow

1. Check prerequisites: weave --version, weave doctor
2. Understand goals: What data? What queries? What scale?
3. Recommend VDB based on goals, infra, budget
4. Create config: weave config create --env
5. Verify: weave health check, weave doctor --fix
6. Hand off to weave-ingest skill for data loading