RANK.AI DATA / LATEST PUBLICATION

AI Agent framework activity

8 software projects ranked by npm downloads (trailing 30 days).
MCP TypeScript SDK is out in front at 167.53M downloads.

8 software projectschecked Jul 27, 2026, 4:22 AM UTCdaily1 failed refreshhow we measure this →CSV / JSON

Follow changes
In front today

MCP TypeScript SDK

167.53M downloadsnpm downloads (trailing 30 days)

Ahead of second place
99.4M
Held by the top three
91%
Middle of the board
7.8M
Below half the leader
7 of 8
Highest 167.53M downloadsMiddle 7.8M
1. MCP TypeScript SDK — 167.53M downloads2. Vercel AI SDK — 68.14M downloads3. LangGraph.js — 10.73M downloads4. LangChain.js — 10.45M downloads5. OpenAI Agents SDK — 5.08M downloads6. Mastra — 4.64M downloads7. Cloudflare Agents SDK — 4.27M downloads8. MCP Inspector — 779.12K downloads
Rank 1one bar per published rowRank 8
WHY THIS RANK / VERIFIED SNAPSHOT

OpenAI Agents SDK

Ranks a fixed agent/MCP project cohort by exact trailing npm downloads with separate GitHub activity context.

Some explanation inputs are unavailable Missing: comparablePriorRanking.

Published rank
#5
Score
5,075,266 downloads
Sample size
5

100% confidence · 100% component coverage · as of Jul 27, 2026, 4:22 AM UTC

SCORE CONSTRUCTION

Component ledger

5/5 evidenced

All components available · 1 evidence record each

Components contributing to OpenAI Agents SDK's rank
ComponentValueWeightContribution
github days since last push00%0
github forks count8720%0
github stars count34710%0
npm downloads trailing 30d5075266100%5,075,266
npm downloads trailing 7d13408790%0
WHY IT MOVED

new since prior snapshot

No rank delta

no_prior_published_snapshot

No component moved since the prior snapshot.

PRIMARY EVIDENCE

Source trail

5 public records
  • Days since last pushgithub days since last push · observed Jul 27, 2026, 4:16 AM UTC0 days100% confidence
    Source
    GitHub REST API
    Grade
    A
    Freshness
    fresh
    Locator
    {"fields":["stargazers_count","forks_count","open_issues_count","pushed_at"],"nodeId":"R_kgDOOzfwkA","endpoint":"/repos/openai/openai-agents-js","repositoryId":993521808}
    Open primary evidence ↗
  • GitHub forksgithub forks count · observed Jul 27, 2026, 4:16 AM UTC872 forks100% confidence
    Source
    GitHub REST API
    Grade
    A
    Freshness
    fresh
    Locator
    {"fields":["stargazers_count","forks_count","open_issues_count","pushed_at"],"nodeId":"R_kgDOOzfwkA","endpoint":"/repos/openai/openai-agents-js","repositoryId":993521808}
    Open primary evidence ↗
  • GitHub starsgithub stars count · observed Jul 27, 2026, 4:16 AM UTC3,471 stars100% confidence
    Source
    GitHub REST API
    Grade
    A
    Freshness
    fresh
    Locator
    {"fields":["stargazers_count","forks_count","open_issues_count","pushed_at"],"nodeId":"R_kgDOOzfwkA","endpoint":"/repos/openai/openai-agents-js","repositoryId":993521808}
    Open primary evidence ↗
  • npm downloads (trailing 30 days)npm downloads trailing 30d · observed Jul 26, 2026, 11:59 PM UTC5,075,266 downloads100% confidence
    Source
    npm Official Downloads API and Registry
    Grade
    A
    Freshness
    fresh
    Locator
    {"windows":{"trailing7d":["2026-07-20","2026-07-26"],"trailing30d":["2026-06-27","2026-07-26"]},"registryEndpoint":"/{package}/latest","downloadsEndpoint":"/downloads/point/{period}/{package}"}
    Open primary evidence ↗
  • npm downloads (trailing 7 days)npm downloads trailing 7d · observed Jul 26, 2026, 11:59 PM UTC1,340,879 downloads100% confidence
    Source
    npm Official Downloads API and Registry
    Grade
    A
    Freshness
    fresh
    Locator
    {"windows":{"trailing7d":["2026-07-20","2026-07-26"],"trailing30d":["2026-06-27","2026-07-26"]},"registryEndpoint":"/{package}/latest","downloadsEndpoint":"/downloads/point/{period}/{package}"}
    Open primary evidence ↗
REPRODUCIBILITYAI agent framework activity methodology · v1-agent-frameworks-mcp-v1-2026-07-2740 observations · 2 sources · passed quality
Snapshot
0c5c895c-1315-4113-aa50-2d66bd46055d
Data hash
b5ab6ea0ad75cd05ebed3543ba15065e9977b43086454313267eb9e7464d8df0
Method hash
82cbe98a11edd94018025fe2f238abf337e857be434ce4820e4755e99929498a

More in AI agents

Task success, cost, speed, reliability, tool use, and human intervention.

3 live tables