Upgrading to v0.7.0

Axiom Wiki v0.7.0 introduces Hybrid Semantic Search (Lexical + Vector) using the Orama engine. This upgrade allows your AI agents to find relevant context through conceptual understanding rather than just keyword matching.

Note

This guide applies to all v0.7.x releases (v0.7.0, v0.7.1, v0.7.2).

🔄 Upgrade Steps (v0.6.0 → v0.7.0)

If you are an existing user upgrading from v0.6.0 to v0.7.0, follow these steps to enable the new semantic capabilities:

1. Configure Embeddings

Run the new interactive setup wizard:

axiom-wiki embed --setup

2. Pick a Provider

Choose the embedding provider that best fits your workflow:

• Google Gemini (Recommended): High performance, generous free tier. Uses text-embedding-004.
• OpenAI: Industry standard. Uses text-embedding-3-small.
• Ollama: 100% local and private. Uses nomic-embed-text.

3. Build Initial Index

The wizard will automatically trigger a full re-index of your existing pages. This process:

1. Generates vector embeddings for every page in your wiki.
2. Stores them in a new local index file: axiom/wiki/search.index.
3. Marks all pages as “vector synced” in your state management.

🚀 What’s New in v0.7.0?

Hybrid Search (Keyword + Vector)

Axiom now uses Reciprocal Rank Fusion (RRF) to merge results from two different search engines:

• Lexical Search: Traditional keyword matching for exact terms and technical identifiers.
• Vector Search: Semantic matching for conceptual queries and intent.

Agent Intelligence Boost

Tools like plan_with_wiki and search_wiki are now “semantic-aware.” Agents can now find relevant documentation even if they use different terminology than what is written in your markdown files.

Real-time Maintenance

The semantic index is kept fresh automatically. Every time you (or your agent) modify a page via autowiki, sync, or the MCP notify_code_change tool, the vector embedding is updated in the background.

🛠 Troubleshooting

Dimension Mismatch

If you switch embedding providers (e.g., from OpenAI to Ollama), the index will detect a dimension mismatch. Simply run:

axiom-wiki embed --reindex

Checking Status

You can check the health of your semantic index at any time:

axiom-wiki status
# OR
axiom-wiki embed --status