Medplum MCP
Securely access and manage FHIR healthcare data stored in Medplum.
Real-time conversation health for AI agents: drift, desync, causal collapse, plus mission clocks.
作者提供的原文
列表按使用分(满分 100)排序,由以下部分组成
| 热度 | 0/80 | GitHub 星标或每周下载量:较大的计 60,较小的计 20 |
|---|---|---|
| 新鲜度 | 12/12 | 30 天内有发布或代码更新得满分 |
| 仍在响应 | 8/8 | 上次检查时云端端点有响应 |
| 使用分 | 20/100 |
质量: 基础检查通过 3/5 项(仅展示,不用于排序) · 上升势头: 暂无数据 (只用于“上升中”标签)
选择你使用的 AI 应用,然后复制设置。
用一句话让 AI 帮你安装
请帮我安装 Horizon Fidelity Monitor。按照 https://mcp.popnx.com/i/leocelis-horizon-fidelity-monitor.md 操作,先把命令给我看,等我同意后再运行。粘贴到能在你电脑上工作的 AI(如 Claude Code、Cursor)。它会读取我们的步骤,先给你看命令,等你同意。无需密钥,无需注册。 · 查看 AI 会读取的全文
通过 3/5 项检查 · 你的提问会经过开发者的服务器(horizon.leocelis.com) · ✓ 命令已检查 2026年11月20日
第一步:你用哪个 AI?
选择你用的应用,步骤会显示在下方。
可连接 Claude · 约 5 分钟
需要 Horizon Fidelity Monitor 的密钥(在开发者网站创建) · 在 claude.ai 或桌面应用设置一次,即可在聊天中使用 · 免费账户可添加 1 个 · Team/Enterprise 方案由所有者(Owner)添加
打开开发者网站,注册账户并创建 API 密钥,下一步要用。价格和免费方案请看开发者网站。
CustomizeConnectors
把这里的内容复制到 Claude 中同名的输入框,然后点击 Continue
Horizon Fidelity Monitorhttps://horizon.leocelis.com/sse在 Request headers 下添加这个名称和值
AuthorizationBearer <在这里粘贴你的密钥>只把密钥粘贴到应用的输入框里,不要粘贴到聊天中或发给任何人
在 Claude 检测到的登录设置页点击 Continue · Authentication 按作者说明选择(Sign in now、Sign in when needed 或 No sign in),然后点击 Add
+ConnectorsHorizon Fidelity Monitor
用 Horizon Fidelity Monitor 告诉我你能帮我做什么,然后做一件给我看✓ 如果 Claude 说正在使用 Horizon Fidelity Monitor,或回答中出现 Horizon Fidelity Monitor 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 Anthropic 帮助页面核实 · 你的界面可能因版本略有不同
ChatGPT 暂时无法连接 Horizon Fidelity Monitor
Horizon Fidelity Monitor 连接时需要你自己的密钥,但 OpenAI 帮助页面只说明了 OAuth 登录,没有我们能核实的填密钥方式
A · 使用能连接 Horizon Fidelity Monitor 的应用
B · 诊所与健康 类中能连接 ChatGPT 的其他服务器
从同类中使用较多的里挑选,仅限用地址即可连接、无需密钥的
可连接 Claude Code · 约 5 分钟
需要 Horizon Fidelity Monitor 的密钥(在开发者网站创建)
打开开发者网站,注册账户并创建 API 密钥,下一步要用。价格和免费方案请看开发者网站。
Mac:按 Command+空格,输入 Terminal · Windows:打开 PowerShell
claude mcp add --transport sse --scope user horizon-fidelity-monitor https://horizon.leocelis.com/sse --header 'Authorization: Bearer {api_key}'按 Enter 前,把“your-value”改成你的密钥
只把密钥粘贴到应用的输入框里,不要粘贴到聊天中或发给任何人
应能看到 Horizon Fidelity Monitor 在列表中并显示已连接
用 Horizon Fidelity Monitor 告诉我你能帮我做什么,然后做一件给我看✓ 如果 Claude Code 说正在使用 Horizon Fidelity Monitor,或回答中出现 Horizon Fidelity Monitor 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 Anthropic 帮助页面核实 · 你的界面可能因版本略有不同
可连接 Cursor · 约 5 分钟
需要 Horizon Fidelity Monitor 的密钥(在开发者网站创建)
打开开发者网站,注册账户并创建 API 密钥,下一步要用。价格和免费方案请看开发者网站。
在 Horizon Fidelity Monitor 的设置中,Authorization 还是空的,填入你的密钥并保存
只把密钥粘贴到应用的输入框里,不要粘贴到聊天中或发给任何人
用 Horizon Fidelity Monitor 告诉我你能帮我做什么,然后做一件给我看✓ 如果 Cursor 说正在使用 Horizon Fidelity Monitor,或回答中出现 Horizon Fidelity Monitor 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 Cursor 帮助页面核实 · 你的界面可能因版本略有不同
可连接 VS Code · 约 5 分钟
需要 Horizon Fidelity Monitor 的密钥(在开发者网站创建)
打开开发者网站,注册账户并创建 API 密钥,下一步要用。价格和免费方案请看开发者网站。
在 Horizon Fidelity Monitor 的设置中,Authorization 还是空的,填入你的密钥并保存
只把密钥粘贴到应用的输入框里,不要粘贴到聊天中或发给任何人
用 Horizon Fidelity Monitor 告诉我你能帮我做什么,然后做一件给我看✓ 如果 VS Code 说正在使用 Horizon Fidelity Monitor,或回答中出现 Horizon Fidelity Monitor 的工具名称,就表示连接成功
菜单名称已于 2026年11月11日 对照 VS Code 帮助页面核实 · 你的界面可能因版本略有不同
Claude Code
claude mcp add --transport sse --scope user horizon-fidelity-monitor https://horizon.leocelis.com/sse --header 'Authorization: Bearer {api_key}'Cursor (~/.cursor/mcp.json)
{
"mcpServers": {
"horizon-fidelity-monitor": {
"url": "https://horizon.leocelis.com/sse",
"headers": {
"Authorization": "Bearer {api_key}"
}
}
}
}VS Code (mcp.json)
{
"servers": {
"horizon-fidelity-monitor": {
"type": "sse",
"url": "https://horizon.leocelis.com/sse",
"headers": {
"Authorization": "Bearer {api_key}"
}
}
}
}说明为自动翻译,变量名保持原样;不确定时请查看原文。
这些设置是根据作者在注册中心的条目生成的。如果 README 的写法不同,请以 README 为准。
依据这个服务器的连接方式和各应用自己的文档,检查于 2026年11月10日。
通过互联网连接,适用于
ClaudeGooseLM StudioMsty StudioCherry StudioCursorVS Code + GitHub CopilotVisual StudioWindsurf (Devin Desktop)ZedKiroClineContinueJetBrains AI AssistantClaude CodeCodex CLIGemini CLIOpenCodenodecodeWarpLibreChatOpen WebUIRaycast
服务器要求先登录,才会列出它的工具。
我们检查的是任何人都能看到的信息;我们没有阅读代码。不适用的检查项不计入总数。 · 检查于 2026年11月20日
注册名称已通过 GitHub 账户 leocelis 验证,该账户拥有此仓库
代码最后更新于3天前
使用 MIT
拒绝访问(HTTP 401),且未提供 OAuth 登录。可能需要密钥,也可能是拦截机器人,尚未确认。
需要你自己复制粘贴密钥或令牌
尚未检查;我们会在检查端点时读取工具列表
如果你是作者,可把这段代码贴到 README。徽章会链接回本页。
[](https://mcp.popnx.com/en/servers/leocelis-horizon-fidelity-monitor)提问、分享技巧或报告无法使用。按“有帮助”数量排序。
还没有评价
与注册中心里这个分类中所有活跃的服务器比较。只有当发布者拥有该仓库时,星标数才计入。
它在 5 项适用的基础检查中通过了 3 项。我们没有阅读代码。在添加密钥或令牌之前,请先看看这个服务器要求哪些权限,并使用只具备所需权限的密钥。
你需要填写:Authorization
它在作者的云端运行。通过互联网连接,无需安装,但你发送的内容会经过作者的服务器。
GitHub 仓库最后更新于3天前,最新版本(0.3.2)于12天前进入注册中心。
支持 MCP 的应用,例如 Claude Desktop、Claude Code、Cursor 和 VS Code。它通过 URL 连接,因此也能用于接受远程 MCP 的应用,例如网页版 ChatGPT(Business 及以上为完整 MCP;Pro 为只读)。
同一分类中有:Medplum MCP, Mirobody, Codescene MCP Server, Marrow, WHOOP — MissingMCP
同一分类、其他作者、命令已通过检查。官方的排在前面,再按星标排序。
如何看卡片:↓ 每周下载量i★ GitHub 星标i三大支柱:评价 · 检查 · 使用情况i更新 = 最新版本发布到注册中心的时间i
Securely access and manage FHIR healthcare data stored in Medplum.
Lab results to LOINC codes and UCUM units, symptoms and diagnoses to ICPC-3. Offline, no API key.
An MCP server that provides CodeScene Code Health analysis tools.
Marrow Health: your Apple Health data over MCP, from the free iOS app. Reads, and writes on demand.
WHOOP recovery, strain, sleep and workouts in Claude via official WHOOP OAuth. Free, open source.
| 本项 Horizon Fidelity Monitor | Medplum MCP | Mirobody | Codescene MCP Server | Marrow | WHOOP — MissingMCP | |
|---|---|---|---|---|---|---|
| 运行位置 | 在云端运行,无需安装 | 在云端运行,无需安装 | 需要安装在电脑上 | 需要安装在电脑上 | 需要安装在电脑上 | 在云端运行,无需安装 |
| 需要密钥 | 是 | 否 | 否 | 否 | 否 | 否 |
| 可在 Claude 中按 + 添加 | 否 | 否 | 否 | 否 | 否 | 否 |
| 基础检查 | 3/5 | 4/4 | 3/3 | 2/3 | 3/3 | 2/4 |
| 许可证 | MIT | Apache-2.0 | Apache-2.0 | NOASSERTION | MIT | MIT |
| 价格 | 登记中未列出费用 | 登记中未列出费用 | 登记中未列出费用 | 登记中未列出费用 | 登记中未列出费用 | 登记中未列出费用 |
按原文从作者的仓库复制。PopNx MCP 没有审查或背书其内容。HTML、图片和脚本已被移除。链接会离开本站。
"Quality is not a model property — it is a conversation property."
Horizon is a real-time conversation health monitor for AI agents. It tracks the structural dynamics of multi-turn conversations — semantic drift, information gain, ontological gap width, temporal desynchronisation, circadian cognitive load, conversation velocity, and causal reachability — dimensions that LLMs do not reliably surface from inside the conversation.
Horizon ships two measurement planes. The conversation plane (above, always present) measures the health of a dialogue turn by turn. The optional mission plane — *Memento Mori* — measures elapsed calendar time against goals: ages, deadlines, stalls, per-entity latency, and share of a finite horizon. It is inert until you configure a store. See Mission plane.
Horizon is not a manipulation, sycophancy, or human-influence detector — it measures conversation *dynamics*, not whether an agent is steering or flattering the user. See LEGAL.md §1.
Why an external monitor? LLMs have *limited and unreliable* self-knowledge: introspection research shows partial self-access that is brittle and degrades on complex or out-of-distribution tasks (Binder et al. 2024; arXiv:2512.12411). So rather than depend on a model reporting its own conversation dynamics, Horizon measures them externally with cheap, deterministic, always-on arithmetic over a small local embedding model, with no LLM calls.
Multi-turn AI agents lose accuracy. The ICLR 2026 oral paper "LLMs Get Lost In Multi-Turn Conversation" (Laban, Hayashi, Zhou & Neville) reports an average 39% performance drop across six generation tasks when instructions arrive over multiple turns — a structural property that standard observability tools (LangSmith, RAGAS, DeepEval) cannot see because they measure responses, not conversations.
Horizon was built to close that gap. It is observability first: it surfaces conversation dynamics that response-level tools miss, using cheap deterministic arithmetic and a small local embedding model, with no LLM calls. In four controlled A/B scenarios where Horizon events drove a re-grounding intervention we measured a +15.7% composite quality lift and 87% fewer hallucination events — but those are *synthetic, scripted scenarios with a hand-tuned controller*, not a production result. Treat them as promising in-house evidence, not a guaranteed outcome (see Validation and LEGAL.md §5). Every signal — information gain, divergence, estimated ontological gap width, causal reachability — is a standard information-theory or arithmetic measure computed on text embeddings and timestamps; see 4D Spacetime Signals for the full definitions.
docs/content/naming-the-category-conversation-dynamics-monitoring.mddocs/content/why-every-production-agent-needs-conversation-dynamics-monitoring.mdThree paths — pick the one that fits your workflow:
The fastest way to add Horizon to any Cursor, VS Code, or Claude Desktop workspace. No Python required.
Request an alpha key → open a Discussion, then add the config for your client:
Cursor (~/.cursor/mcp.json):
{
"mcpServers": { "horizon": { "url": "https://horizon.leocelis.com/sse", "headers": { "Authorization": "Bearer YOUR_KEY_HERE" } } } }
VS Code / GitHub Copilot (.vscode/mcp.json in your workspace):
{
"servers": { "horizon": { "type": "http", "url": "https://horizon.leocelis.com/sse", "headers": { "Authorization": "Bearer YOUR_KEY_HERE" } } } }
VS Code note: Use
"servers"(not"mcpServers") and"type": "http"— VS Code tries Streamable HTTP first and falls back to SSE automatically, so"type": "http"works with the/sseURL.
Claude Desktop (~/Library/Application Support/Claude/claude_desktop_config.json):
{
"mcpServers": { "horizon": { "url": "https://horizon.leocelis.com/sse", "headers": { "Authorization": "Bearer YOUR_KEY_HERE" } } } }
That's it. Reload your MCP client and three tools appear: new_conversation, process_turn, configure_session. With a mission store configured, seven more appear: clock_register, clock_progress, clock_status, clock_propose, clock_ack, clock_close, associate_mission (see Mission plane).
Alpha access: Horizon's hosted endpoint is in private alpha. Keys are distributed to agent developers who want to monitor real projects. Open a Discussion to request one — describe your use case and we'll send a key.
Python 3.10–3.14.
pip install horizon-monitor
Verify your install (exercises the full pipeline on 5 canonical scenarios, ~25s):
horizon-validate
git clone https://github.com/leocelis/horizon && cd horizon
pip install -e '.[mcp]' horizon serve # stdio — for Cursor, Claude Desktop horizon serve --transport sse --port 3847 # SSE — for web/team deployments
Add to ~/.cursor/mcp.json:
{
"mcpServers": { "horizon": { "command": "horizon", "args": ["serve"] } } }
Full Cursor and Claude Desktop setup guides: docs/integrations/
Standard observability tools evaluate individual response quality. Horizon evaluates conversation quality — a structurally different problem:
| Tool | What it sees | What it misses |
|---|---|---|
| LangSmith, Braintrust | Latency, cost, per-response quality | Deterministic, every-turn structural signals |
| RAGAS, DeepEval | Faithfulness, relevance per turn (DeepEval also has sampled multi-turn LLM-judge metrics) | Zero-LLM-call, real-time scoring on every turn |
| Langfuse, Arize Phoenix | Session-level LLM-judge evaluation | Deterministic, always-on scoring at sub-50ms |
| Human raters | Subjective quality | Systematic structural decay |
| Horizon | Conversation dynamics | Intentionally nothing |
Horizon does not replace per-response or LLM-judge quality tools. The differentiator is *how* it measures: deterministic, zero-LLM-call arithmetic on every single turn — effectively free and always-on — versus the alternative of sampled LLM-judge evaluations, which cost per sample and typically run offline or async rather than in real time.
from horizon_monitor import FidelityMonitor
from datetime import datetime, timezone
monitor = FidelityMonitor() session_id = monitor.new_conversation(metadata={"domain": "technical"})
result = monitor.process_turn( session_id, human_message="How does Python handle memory management?", agent_response="Python uses reference counting and a cyclic garbage collector...", timestamp=datetime.now(timezone.utc).isoformat(), )
print(f"Fidelity: {result.fidelity_score:.2f}") print(f"Health: {result.health_status}") print(f"Circadian factor: {result.circadian_factor:.2f}") print(f"Causal horizon: {result.reachable_turns} reachable turns") for event in result.events: print(f" Event: {event.type} (confidence={event.confidence:.2f})")
from openai import OpenAI
from horizon_monitor import FidelityMonitor
monitor = FidelityMonitor() session_id = monitor.new_conversation() client = monitor.wrap(OpenAI(), session_id)
response = client.chat.completions.create( model="gpt-4o", messages=[{"role": "user", "content": "Tell me about quantum computing."}] )
traj = monitor.get_trajectory(session_id) print(f"Fidelity: {traj.current_fidelity:.2f} T*: {traj.estimated_t_star}")
monitor.wrap() accepts custom timestamp and context providers for testing and replay.
from anthropic import Anthropic
from horizon_monitor import FidelityMonitor
monitor = FidelityMonitor() session_id = monitor.new_conversation() client = monitor.wrap(Anthropic(), session_id)
response = client.messages.create( model="claude-3-5-sonnet-20241022", max_tokens=1024, messages=[{"role": "user", "content": "Explain RLHF."}] )
from langchain_openai import ChatOpenAI
from horizon_monitor import FidelityMonitor from horizon_monitor.integrations.langchain import HorizonCallback
monitor = FidelityMonitor() session_id = monitor.new_conversation() callback = HorizonCallback(monitor, session_id)
llm = ChatOpenAI(callbacks=[callback]) llm.invoke("Explain the CAP theorem.") print(f"Fidelity: {callback.last_result.fidelity_score:.2f}")
from agents import Agent, Runner
from horizon_monitor import FidelityMonitor
monitor = FidelityMonitor() session_id = monitor.new_conversation() agent = Agent(name="assistant", model="gpt-4o-mini", instructions="You are helpful.")
for user_message in conversation: result = Runner.run_sync(agent, user_message) monitor.process_turn(session_id, human_message=user_message, agent_response=result.final_output, timestamp=datetime.now(timezone.utc).isoformat())
"Spacetime" here is a metaphor, not physics. The relativity vocabulary (Minkowski interval,
light cone, proper time) is *design inspiration* — it shaped which quantities we compute. Every
signal below reduces to a standard information-theory or arithmetic measure on text embeddings and
timestamps, listed in the Plain definition column. Nothing in Horizon's behavior or validation
depends on the analogy being literally true, and the Lorentzian
interval_classis emitted as
descriptive metadata only — no event or score depends on it.
Every process_turn() returns a TurnResult with 33 fields across six signal families:
| Signal | Description |
|---|---|
fidelity_score | Composite conversation health [0, 1] |
igt_value | Information Gain per Turn — semantic novelty |
divergence_score | Jensen-Shannon proxy for intent/response gap |
twr_value | Token Waste Ratio — semantic redundancy |
consistency_score | Bipredictability — structural coherence |
epsilon_t | Estimated ontological gap width [0, 1] |
health_status | healthy / degrading / critical / converged |
conversation_mode | execute / explore / refine / learn (auto-detected) |
timestamp)| Signal | Description |
|---|---|
gap_seconds | Wall-clock gap since last turn |
estimated_retention | Human memory retention (Ebbinghaus half-life model) |
| circadian_factor | Human cogn