LoginSign Up
← Back to all sparks
A

Apache Answer

COLLAB
Velocity2.5

Self-hosted Q&A community platform

answer.apache.org ↗

Apache Answer is turning its Q&A engine into a searchable AI knowledge base, one RC at a time.

q-and-asemantic-searchai-chataccess-controlself-hosted
◆Current state
Apache Answer ships through repeated release candidates: v2.0.1 added semantic search in AI chat backed by a vector-search plugin, v2.0.2 kept model reasoning output in chat history and let admins skip email verification, and v2.0.3-RC1 moved the frontend from Create React App to Vite. Every cut also carries a long list of authorization fixes, many touching MCP content access.
◆Where it's heading
The product is layering AI retrieval on top of its community Q&A data while hardening who can see what, including through the MCP surface. The Vite migration and Helm persistence options point to a project paying down build and deployment debt ahead of further feature work.
◆Prediction
Expect v2.0.3 to finalize after one or two more RCs, followed by more vector-search and AI chat refinements; the cadence suggests a new minor line roughly every two months.

◆Recent moves

  1. 8d ago

    v2.0.3 RC1: Vite frontend, Helm volume reuse, authorization sweep

    The 2.0.3 line swaps Create React App for Vite, lets Helm installs reuse an existing volume, and only exposes semantic search in AI chat when a vector plugin is actually installed. A broad sweep of visibility and authorization checks, including MCP content access, continues the security hardening seen in every recent cut.

    View source ↗
  2. 3mo ago

    v2.0.2 RC3 fixes question editing with short links

    RC3 adds one fix on top of RC2: editing a question no longer fails when short links are enabled.

    View source ↗
  3. 3mo ago

    v2.0.2 RC2 adds a missing dependency license entry

    RC2 only adds a missing license entry for a bundled dependency; the feature set is unchanged from RC1.

    View source ↗
  4. 3mo ago

    v2.0.2 RC1: AI chat keeps reasoning, optional email verification

    v2.0.2 keeps model reasoning output across AI chat follow-ups, lets operators disable mandatory email verification, and moves new-question email throttling to its own worker. Most of the remaining notes are permission fixes for API keys, revision audits and MCP retrieval.

    View source ↗
  5. 4mo ago

    v2.0.1 RC2 re-cut of the semantic search release

    RC2 republishes the v2.0.1 notes with no new user-facing changes listed beyond RC1.

    View source ↗
  6. 5mo ago

    v2.0.1 RC1 adds semantic search to AI chat via vector plugin

    v2.0.1 is where AI chat starts answering from the site's own content: a vector-search plugin embeds questions and answers, and chat can query them semantically. It also fixes the attachment uploads that v2.0.0 broke and tightens API key scopes.

    View source ↗