Git.Top 2.0 Roadmap

The Knowledge Graph of Open Source

Git.Top 2.0 is mostly through the first product cycle: the Worker-only surface now exposes human pages and agent-readable APIs for discovery, alternatives, comparison, graph reasoning, scoring, Atlas, workflow, and trust.

Overall completion88/100

Average progress across the six 2.0 product phases.

Implemented phases4

Core public surfaces that are live on the Worker.

Active phases2

Areas still being deepened through iteration.

Worker surface1

Worker remains the only product surface.

Current Focus

What continues next

Expand high-intent topic pages and product guides.
Improve Atlas exploration depth without reintroducing a separate frontend surface.
Reduce data trust risk through sync freshness, classification review, and stronger alternatives evidence.
Keep REST, MCP, OpenAPI, llms.txt, Agent Map, and human pages aligned.
1 implemented

92/100

homepage-entry-points

Homepage And Entry Points

Make the first visit explain Git.Top's value as the knowledge layer for open source.

Shipped

Homepage hero states The Knowledge Graph of Open Source.
Discover, Compare, Alternatives, Workflow, Agent API, Score, Trust, Graph, Atlas, Recommend, and Ecosystems entry cards are live.
High-value topic pages cover agents, MCP, RAG, browser automation, AI IDE, comparison, graph, score, workflow, alternatives, and trust.

Next

Continue adding topic pages for high-intent discovery and SEO clusters.

APIs And MCP

2 implemented

88/100

project-knowledge-graph

Project Knowledge Graph

Make relationship pages the core moat: related projects, alternatives, dependencies, deployments, and graph JSON.

Shipped

Project graph pages expose overview, relationship groups, deployment targets, dependencies, related projects, alternatives, filters, and exploration paths.
Graph APIs expose summary, graph stats, relationship groups, nodes, edges, and next actions.
Graph guide pages explain how humans and agents should read project relationships.

Next

Introduce a client-side graph library only when relationship density needs richer interaction.
3 implemented

86/100

alternatives-engine

Alternatives Engine

Turn alternatives into an SEO and decision-support entry point.

Shipped

Alternatives index and project alternatives pages are live.
Alternative matches include fit summary, adoption notes, replacement risk, comparison links, and next actions.
Matching uses explicit alternatives, category, deployment, use cases, topics, dependencies, language, maintenance, and Cloudflare readiness.

Next

Improve semantic similarity when vector infrastructure is introduced.
4 implemented

90/100

project-score

Git.Top Score

Stop relying on GitHub stars as the primary quality proxy.

Shipped

Project pages and score pages show Git.Top Score, Agent Score, Quality Score, weighted dimensions, confidence, evidence, and risk flags.
Score APIs expose dimensions, strongest and weakest dimensions, adoption guidance, related scores, next actions, and score confidence.
Quality and trust pages separate release score from corpus data trust.

Next

Keep improving confidence evidence and reviewed overrides for flagship projects.
5 active

82/100

atlas

Atlas

Create an exploratory map for open-source ecosystems.

Shipped

Atlas maps are live for Cloudflare, AI Agents, MCP, RAG, and Browser AI.
Atlas now includes AI IDE, LLM Gateway, AI Observability, and Workflow Automation ecosystems for denser exploration.
Atlas exposes stats, exploration paths, graph nodes, graph edges, representative projects, and journeys.
Atlas Journeys is live as a human page and JSON API for turning ecosystem maps into ordered exploration routes.
Atlas now exposes ecosystem-specific comparison paths so users and agents can move from a map to a decision matrix.
MCP get_atlas exposes the same curated maps to agents.

Next

Add richer interactive graph controls when the Worker UI needs dense exploration.
6 active

89/100

agent-api

Agent API

Make Git.Top a default structured knowledge source for agents.

Shipped

REST endpoints cover project lookup, recommendation, workflow, alternatives, graph, compare, score, trends, atlas, journeys, quality, quickstart, recipes, and agent map.
MCP tools map to the same product concepts and support D1-required strict mode.
llms.txt, llms-full.txt, OpenAPI, schema, Agent Map, Quickstart, Recipes, Examples, and workflow pages make the product machine-discoverable.
Quickstart and Recipes now include Atlas journeys and comparison paths as first-class agent workflows.
Knowledge Graph API and MCP integration guide pages now explain the human page, REST, MCP, OpenAPI, LLM discovery, output field, and trust-field pairs.
API Examples now exposes copyable REST, MCP, and GRP calls through both /examples and /api/examples.
Trust Gate is live at /trust and /api/trust to combine D1 source, sync freshness, release score, data trust score, and risk level into one production-readiness decision.

Next

Continue tightening OpenAPI schemas and MCP integration recipes for common production workflows.

Next Milestones

Remaining product work

Add more decision-oriented topic pages for SEO and onboarding.
Add richer Atlas journeys and ecosystem-specific comparison paths.
Improve production data trust by shrinking stale sync and weak alternatives review queues.
Extend MCP integration recipes for common agent selection workflows.