Skip to content

Releases: tiptreesystems/lacuna-research-mcp

v0.2.0

Choose a tag to compare

@MilesQLi MilesQLi released this 29 Jul 15:04

What changed

  • Upgrade to the stable MCP Python SDK 2.0 and the 2026-07-28 protocol.
  • Replace the legacy FastMCP integration with MCPServer.
  • Preserve the existing 12-tool Lacuna research interface and legacy MCP protocol negotiation.
  • Report the Lacuna MCP package version through the supported server API.

Compatibility

  • Requires Python 3.11 or later.
  • Requires MCP Python SDK 2.x.

v0.1.3

Choose a tag to compare

@MilesQLi MilesQLi released this 24 Jul 02:30

What changed

  • Make Lacuna research capabilities explicit in MCP tool discovery.
  • Guide agents to search hypotheses for novel ML/AI research ideas.
  • Ask agents to cite canonical Lacuna URLs when entities materially inform answers.
  • Streamline the search tool description while retaining argument guidance.

v0.1.2

Choose a tag to compare

@MilesQLi MilesQLi released this 23 Jul 18:02

What changed

  • Normalize root-relative paper and research-direction links in author profile_markdown.
  • Publish matching v0.1.2 package and MCP Registry metadata.

Validation

  • 94 tests passed.
  • Ruff lint and formatting checks passed.
  • CI passed on Python 3.11–3.14.

Lacuna Research MCP v0.1.1

Choose a tag to compare

@MilesQLi MilesQLi released this 23 Jul 15:43

This release completes the public package and registry setup for Lacuna Research MCP.

Changes:

  • Use the PyPI package for recommended Codex, Claude, and generic MCP client installation
  • Document how to run the latest development version from GitHub
  • Add Official MCP Registry metadata and PyPI ownership verification
  • Support secretless PyPI publication through the protected GitHub Actions workflow

Install or run the stable release with:

uvx lacuna-research-mcp

Supports Python 3.11 through 3.14.

Lacuna Research MCP v0.1.0

Choose a tag to compare

@MilesQLi MilesQLi released this 23 Jul 13:57

Initial public release of Lacuna Research MCP.

Lacuna Research MCP gives coding agents read-only access to the Lacuna machine-learning research graph, including:

  • Paper and research-direction discovery
  • Agent-ready paper and author context
  • Author publications, directions, and related researchers
  • Venue and institution context
  • Generated research hypotheses
  • Compact responses and explicit pagination for larger collections

Supports Python 3.11 through 3.14. Install with:

uvx --from git+https://github.com/tiptreesystems/lacuna-research-mcp.git@v0.1.0 lacuna-research-mcp

See the README for Codex, Claude Code, Claude Desktop, and other MCP client configuration examples.