Emerging AI agents & MCP servers

Data as of 2026-06-13 · first weekly scoring run pending · updated weekly

Early-launch ranking: ordered by real stars gained over our first 4 days of daily tracking. The full acceleration model described below activates automatically as the dataset matures — nothing here is ranked by total popularity.

A project is emerging when it is young (first tracked within ~6 months), still small (under 5,000 stars), and already gaining 30+ stars a week. These are the projects building real momentum before most people have heard of them. Projects that merely continue a deprecated predecessor are excluded — that's continuation, not emergence.

MCP servers: 20
#ProjectCategoryΔ★ 7dTotal ★
1io.github.butterbase-ai/mcp mcp-server+2112,019
2Codebase Memory mcp-server+2953,404
3TokenSave mcp-server+16197
4yutu mcp-server+32505
5io.github.BetterDB-inc/monitor mcp-server+751,218
6OpenAccountants mcp-server+10184
7OpenOSINT mcp-server+32629
8io.github.mnemox-ai/tradememory-protocol mcp-server+371,213
9TypeUI mcp-server+291,134
10jau123/MeiGen-AI-Design-MCP mcp-server+311,425
11agent-device mcp-server+502,714
12io.github.vkhanhqui/figma-mcp-go mcp-server+191,094
13ai.smithery/ChiR24-unreal_mcp mcp-server+12708
14samuelgursky/davinci-resolve-mcp mcp-server+201,227
15Power BI Modeling MCP Server mcp-server+13856
16Repowise mcp-server+342,285
17Figma MCP Server mcp-server+221,601
18SearXNG Search mcp-server+12894
19io.github.basicmachines-co/basic-memory mcp-server+423,222
20Coding-Solo/godot-mcp mcp-server+534,149

Methodology

Age window
First release or tracking start within roughly the last 6 months.
Size ceiling
Under 5,000 stars — past that, a project is established, not emerging.
Momentum floor
At least 30 stars gained in the last 7 days, so dormant new repos don't qualify.
Succession filter
A project that officially succeeds a deprecated predecessor is excluded — its growth is migration, not discovery.
All signals are computed from daily snapshots of stars, forks, releases, and npm/PyPI downloads, aggregated weekly. Every number above is traceable to a snapshot date — nothing is editorial, and rankings cannot be bought. Full details on the methodology page.