SeenRelay

SeenRelay

Enables AI agents to check and contribute recent observations of source-backed facts, reducing redundant revalidation of unchanged information.

Category
访问服务器

README

SeenRelay

A cooperative freshness cache for AI agents.

Think of SeenRelay as a shared cache for freshness evidence rather than content. A normal cache lets a later request avoid fetching the same bytes again. SeenRelay lets a later check reuse recent source-backed evidence produced by the same integration, the same fleet, or another agent instead of blindly repeating the full validation path.

CHECK before repeating validation work. OBSERVE what you independently found while doing work you already needed to do.

SeenRelay has exactly two domain operations: CHECK and OBSERVE. It reports recent observations, not universal truth.

For application workflows where CHECK must run whenever a selected validation path executes, use the deterministic JavaScript/TypeScript or Python wrapper. MCP remains the standard discovery and model/tool-routing interface. Both paths expose the same two SeenRelay operations; the wrapper changes execution placement, not protocol semantics.

Why connect now

SeenRelay remains useful before broad network coverage exists:

  • with no prior observation, CHECK returns UNKNOWN and your existing workflow continues;
  • when your agent validates the source anyway, OBSERVE can make that result reusable for later callers;
  • later runs in the same integration or fleet can benefit from observations produced by their own normal work;
  • when an observation carries an ETag or Last-Modified validator, CHECK can return it as an explicitly unverified conditional-request hint for cheaper source confirmation;
  • external observations add coverage as the network grows, but they are not required for the integration to function.

Access is currently free and requires no account or API key.

Start here

  • Recommended deterministic application integration: clients/README.md
  • Integration choices and MCP setup: docs/CLIENTS.md
  • Quickstart: docs/QUICKSTART.md
  • Protocol contract: docs/PROTOCOL.md
  • Web quickstart: https://seenrelay.com/quickstart
  • Web client integrations: https://seenrelay.com/clients
  • MCP endpoint: https://seenrelay.com/mcp
  • Official MCP Registry: io.github.ovladon/seenrelay
  • OpenAPI: https://seenrelay.com/openapi.json
  • Machine descriptor: https://seenrelay.com/service.json
  • Machine-oriented index: https://seenrelay.com/llms.txt

How it works

  1. An agent is about to validate a structured source-backed fact.
  2. It sends CHECK with the fact identity and the value it already knows.
  3. SeenRelay returns recent observation status for that exact fact.
  4. If policy still requires validation, the agent performs its normal source check.
  5. It sends OBSERVE with the independently obtained result so later callers can reuse the freshness evidence.

Possible statuses are SAME_OBSERVED, CHANGED_OBSERVED, CONTESTED, STALE, and UNKNOWN.

SAME_OBSERVED means the same value was recently observed for the same deterministic fact identity. It is not a truth verdict. The consuming agent decides whether the evidence is sufficient for its own policy.

Deterministic client path

The reference wrappers are the recommended path when application code must guarantee that CHECK runs before selected validation work. They put SeenRelay directly around the validation the application already performs rather than relying on a model to select the MCP tool.

MCP remains fully supported as the standard discovery and tool interface when model-selected tool routing is appropriate.

The wrappers default to shadow mode, fail open on relay-side failure, retain no completed CHECK result cache, and require an explicit caller policy before reuse can suppress validation. They do not add a SeenRelay domain operation.

Validator-assisted revalidation

OBSERVE can include source metadata such as an etag or last_modified value obtained during the caller's normal validation. When the newest usable observation carries one of those validators, a fresh CHECK can return it together with source_validator_assurance: observer_supplied_unverified and a safe conditional_request_hint.

That hint is an optimization, not a truth claim. If the caller's policy still requires source confirmation, it can attempt If-None-Match or If-Modified-Since before invoking a more expensive browser, extraction, API, or model pipeline. A 304 Not Modified response is confirmation from the source, not from SeenRelay.

Fact identity

SeenRelay uses the versioned seenrelay-fact-v3 identity contract. Identity precedence is:

  1. stable source-native locator (json_pointer, element_id, source_key);
  2. canonical machine predicate when no stable locator exists.

Human-readable subject text and mutable observed content do not enter the fact key. Source URLs are canonicalized deterministically without browsing. Credential- or signature-bearing source URLs are rejected before stateful admission.

See docs/PROTOCOL.md for the complete contract.

Observer provenance

OBSERVE supports optional transport-independent ed25519-v1 proof-of-possession. A valid proof establishes key possession, continuity, and payload integrity. It does not establish legal identity, independent real-world actor identity, or truth.

Access and contribution

SeenRelay issues signed ephemeral operational leases without account creation. CHECK and OBSERVE are currently free to use. Contribution credit is based on later qualifying reuse rather than raw submission volume.

Hive classes describe operational contribution only; they are not identity or truth scores.

Public interfaces

The canonical domain is seenrelay.com.

  • Browser Accept: text/html at / receives the public landing page.
  • Generic/API requests to / receive the machine descriptor.
  • /service.json exposes the explicit machine descriptor, including the deterministic-wrapper and MCP integration paths.
  • /public-stats.json exposes privacy-safe aggregate activity.
  • /openapi.json exposes the REST contract.
  • /mcp exposes MCP 2026-07-28.
  • /quickstart and /clients provide integration instructions.

Product boundary

SeenRelay itself does not browse or search fact sources, perform on-demand verification, call an LLM to decide truth, or expose a shared general-agent memory. UNKNOWN simply means no sufficiently recent reusable observation is available.

Architecture

  • Vercel managed deployment
  • Neon Postgres state store
  • Hono + TypeScript / Node 22
  • REST/OpenAPI
  • MCP 2026-07-28 through the official v2 server SDK
  • vendorable deterministic JavaScript/TypeScript and Python client wrappers
  • authenticated human-only Control Room for runtime operations and incident controls

A2A is monitored but is not advertised as an implemented product interface.

Verification

npm run check performs TypeScript checks, product guardrails, production dependency auditing, structural tests, and runtime tests. The dedicated Client Wrappers workflow executes the JavaScript and Python wrapper regressions. The Preview Release Gate additionally exercises REST, MCP, fact identity, security boundaries, runtime controls, and reuse accounting against the exact Preview deployment SHA before Production promotion.

Builds use the committed lockfile and npm ci.

Maintenance

Dependabot and Standards Watch prepare isolated maintenance work. Production changes remain subject to compatibility, security, CI, and Preview verification gates.

Bootstrap

npm ci
npm run check
npm run db:migrate

Deployment details are in docs/DEPLOYMENT.md.

推荐服务器

Baidu Map

Baidu Map

百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。

官方
精选
JavaScript
Playwright MCP Server

Playwright MCP Server

一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。

官方
精选
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

官方
精选
本地
TypeScript
VeyraX

VeyraX

一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。

官方
精选
本地
graphlit-mcp-server

graphlit-mcp-server

模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。

官方
精选
TypeScript
Kagi MCP Server

Kagi MCP Server

一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

Exa MCP Server

模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。

官方
精选