AgentWyre — AI Ecosystem Intelligence
Query the AgentWyre API for curated, hype-checked intelligence about the AI ecosystem.
When to Use
- - User asks "what's new in AI" or "any breaking changes"
- Checking if dependencies have security vulnerabilities or updates
- Getting the daily AI brief or latest flash signals
- Searching for information about specific AI tools, models, or frameworks
Quick Start
CODEBLOCK0
API Endpoints
| Endpoint | Auth | Description |
|---|
| INLINECODE0 | No | Free tier (full wire, 2-day delay) |
| INLINECODE1 |
API key | Today's full feed |
|
GET /api/feed/{date} | API key | Feed by date |
|
GET /api/flash/latest | Pro key | Latest flash signal |
|
GET /api/status | No | Service status |
|
GET /api/languages | No | Supported languages |
|
GET /api/advisories | Pro key | Security advisories |
|
GET /api/costs | Pro key | AI model pricing |
Using the Helper Script
CODEBLOCK1
Set AGENTWYRE_API_KEY environment variable for authenticated access. Without it, you get the free tier (full wire, 2-day delay).
MCP Server
For deeper integration, install the MCP server:
CODEBLOCK2
Tiers
- - Free — Full daily wire, 2-day delay. No key needed.
- Daily ($2.99/mo) — Same-day access, search, 8 languages.
- Pro ($9.99/mo) — Everything + hourly flash, security advisories, cost tracking.
- USDC on Base accepted — no international bank fees.
Interpreting Signals
Each signal includes:
- - title + summary — What happened
- category — release, security, ecosystem, infrastructure, research, breakingchange
- severity — critical, high, medium, low
- confidence — 1-10 (single source capped at 6)
- hypecheck —
hype_level (verified/promising/overhyped/vaporware), reality, red_flags, INLINECODE12 - action — Recommended steps with commands and rollback instructions
- sources — Verifiable URLs
Always present actions to the user before executing. All actions include requires_user_approval: true.
More info: https://agentwyre.ai/faq
AgentWyre — AI生态系统情报
查询AgentWyre API,获取经过筛选和验证的AI生态系统情报。
使用场景
- - 用户询问AI领域有什么新动态或是否有重大变更
- 检查依赖项是否存在安全漏洞或更新
- 获取每日AI简报或最新闪讯信号
- 搜索特定AI工具、模型或框架的信息
快速开始
bash
获取今日免费信号(无需密钥)
curl -s https://agentwyre.ai/api/feed/free | python3 -m json.tool
获取完整信息流(需要API密钥)
curl -s -H Authorization: Bearer $AGENTWYRE
APIKEY https://agentwyre.ai/api/feed
检查服务状态
curl -s https://agentwyre.ai/api/status
API端点
| 端点 | 认证 | 描述 |
|---|
| GET /api/feed/free | 否 | 免费层级(完整信息流,延迟2天) |
| GET /api/feed |
API密钥 | 今日完整信息流 |
| GET /api/feed/{date} | API密钥 | 按日期获取信息流 |
| GET /api/flash/latest | Pro密钥 | 最新闪讯信号 |
| GET /api/status | 否 | 服务状态 |
| GET /api/languages | 否 | 支持的语言 |
| GET /api/advisories | Pro密钥 | 安全公告 |
| GET /api/costs | Pro密钥 | AI模型定价 |
使用辅助脚本
bash
获取最新信号(免费层级)
python3 scripts/agentwyre.py signals
获取特定语言的信号
python3 scripts/agentwyre.py signals --lang ja
搜索(需要每日/Pro密钥)
python3 scripts/agentwyre.py search langchain breaking change
检查特定包的安全性
python3 scripts/agentwyre.py security openai langchain torch
获取闪讯信号(仅限Pro)
python3 scripts/agentwyre.py flash
设置AGENTWYREAPIKEY环境变量以进行认证访问。未设置时,您将使用免费层级(完整信息流,延迟2天)。
MCP服务器
如需更深度的集成,请安装MCP服务器:
json
{
mcpServers: {
agentwyre: {
command: npx,
args: [agentwyre-mcp],
env: { AGENTWYREAPIKEY: your_key }
}
}
}
层级
- - 免费 — 完整每日信息流,延迟2天。无需密钥。
- 每日($2.99/月) — 当日访问、搜索、8种语言。
- Pro($9.99/月) — 全部功能 + 每小时闪讯、安全公告、成本追踪。
- 接受Base链上的USDC — 无国际银行手续费。
信号解读
每个信号包含:
- - 标题 + 摘要 — 发生了什么
- 类别 — 发布、安全、生态系统、基础设施、研究、重大变更
- 严重程度 — 严重、高、中、低
- 置信度 — 1-10(单一来源上限为6)
- 炒作验证 — 炒作级别(已验证/有前景/过度炒作/空想),实际情况,危险信号,积极信号
- 操作建议 — 推荐步骤,包含命令和回滚说明
- 来源 — 可验证的URL
始终在执行前向用户展示操作建议。所有操作均包含requiresuserapproval: true。
更多信息:https://agentwyre.ai/faq