本页面为机器翻译。阅读英文原文。

Component

core/ai-agent/langflow

The Langflow visual flow-builder, run in-cluster as the image a Release names. Deploying it gives the browser editor and the endpoint serving its flows, on a volume holding them and their encryption key. This component exposes the flow a caller invokes, the per-request tweaks that call may carry, and how the editor is reached: signing in is required by default, so the seeded administrator's password is attached or the build halts, while an Aggregate wanting the editor open to anyone says so. A flow authored in the Aggregate is loaded at each start on both targets, which an image at Langflow 1.5.1 or older does only with signing in off. The catalog pins no image: the Release names it in `infra.image`, and the build halts until it does. Each vector store wired into the optional `vector_store` socket arrives under variables named for the connection's alias: its address as AGGREGA_VECTOR_STORE_<ALIAS>_URL, and the connection's credential as AGGREGA_VECTOR_STORE_<ALIAS>_CREDENTIAL. A connection writing an `inject:` block receives the names it dictates instead, read by a flow as global variables; its credential stays. Each tool wired into the optional `tools` socket, a Model Context Protocol server reached over HTTP, is written into the flow's MCP Tools node named for the connection's alias. With `values.metrics_enabled` true Langflow's metrics are served at /metrics on port 9090, which checks no caller, and published as `metrics_endpoint`. On Kubernetes, reach for `langflow`, the workload synthesis, where the build must fetch nothing or the metrics are wanted, and for `langflow-chart`, rendering Langflow's own chart, which refuses them, where a chart repository is reachable at build time.

版本

1.0.01.1.0

来源

官方由本目录认定为官方发布者的签名身份签名。
内容摘要
sha256:3903e567dc871d167708fe33f309d7560959fd92d2258e3685a7b179b7baf1d0
发布时间
2026-11-10T10:08:13Z
签名身份
https://token.actions.githubusercontent.com/https://github.com/aggrega-ai/aggrega-hub/.github/workflows/catalog-publish-sign.yml@refs/heads/main

签名

  • 官方application/vnd.dev.sigstore.bundle.v0.3+json

    https://token.actions.githubusercontent.com/https://github.com/aggrega-ai/aggrega-hub/.github/workflows/catalog-publish-sign.yml@refs/heads/main

    sha256:fcee870dbce0969a72d4187dec219f78234be7a5ed9cbbcf640c99645f8d0719

此版本的引用

oci://hub.aggrega.ai/component/core/ai-agent/langflow:1.0.0

开发者参数

开发者在 Aggregate 中设置的值。

扩展 BaseComponents#/schemas/BaseAgentParameter

名称类型默认值必需描述
flow_idstring—是—
flow_definitionstring—否A Langflow flow definition (JSON) the developer authors, realized on both targets: the helm target carries it in a ConfigMap mounted as the directory LANGFLOW_LOAD_FLOWS_PATH names, and the compose target in a compose config mounted as one file in that directory, every $ written as $ so Docker Compose delivers it unchanged. Langflow loads the directory into the administrator's flows at each start, whether or not signing in is required; where the definition carries the id of a flow already loaded, that flow is updated in place, so an edit made to it in the editor is replaced at the next start. An image at Langflow 1.5.1 or older loads it only with require_authentication false. Where tools connections are wired, each MCP Tools node whose server is named for a connection's alias has that connection's address, and the header carrying its credential, written into the node's server configuration, every other part of the definition kept and the whole written back as JSON with its keys sorted; with no tools connection the definition is carried as written. A node naming a server no tools connection answers halts the build, as does a tools connection no node names, a node it names that starts a server of its own, and a tools connection wired with no definition. A server of the same name stored in Langflow's settings takes precedence over the configuration written here. An Aggregate's text is evaluated as a template before it is read, so a definition whose component source carries {{ halts the build when written into an Aggregate as it stands; write each {{ in it as {{"{{"}}, and interpolation restores the definition's text. When unset, the platform deploys with no seeded flow.
tweaksobject—否—
require_authenticationbooleantrue否Whether reaching the editor requires signing in. At the default the platform's built-in sign-in bypass is switched off and the seeded administrator account is the only way in, so that account's password must be attached to the admin_password slot; leaving this at the default and attaching nothing is refused at build rather than deployed with an editor nobody can sign in to. Set it to false to run the editor with the bypass on, where everyone who can reach the service arrives already signed in as the administrator and no credential is needed.
admin_userstringadmin否Login name of the administrator account the platform seeds at startup. Read only where the editor requires a sign-in; with require_authentication false no account is seeded and nothing reads this value.
metrics_enabledbooleanfalse否Turns on Langflow's own metrics, in the Prometheus format, on a port of their own, 9090, and publishes their address as `metrics_endpoint`; while false that key is empty. The metrics port checks no caller, so anything that reaches it can read them. The langflow-chart Recipe refuses true, since the chart it renders has no Service port for the metrics; the workload-synthesis and compose Recipes serve them.

运维规格

运维人员按环境调整的基础设施规格。

名称类型默认值必需描述
imagestring—否Container image running the Langflow platform.
source_versionstring—否Version of Langflow's own langflow-ide Helm chart the chart-backed Recipe pulls, defaulting to the version that Recipe was checked against. The workload-synthesis Recipe pulls no chart and reads nothing here.
storageobject—否—

输出契约

此组件通过 uses 拓扑向使用方发布的值。

名称类型默认值必需描述
metrics_endpointstring—否The full address Prometheus scrapes this component's metrics at — scheme, host, port and path, `http://<host>:9090/metrics`, the host being the platform's Service name. Empty while values.metrics_enabled is false.

uses 套接字

此组件用于建立连接的依赖套接字。通过某个套接字接入的每个连接,其目标都必须满足该套接字的标签选择器,且每个连接都可以在该套接字声明的槽位中携带附加项。必需槽位须由 Aggregate 或 Release 通过该套接字接入的每个连接来填充。

model

选择器
component.aggrega.ai/category in (ai-model, gateway), model.aggrega.ai/api
注入
—

经由此套接字的每个连接可填充的槽位

槽位选择器必需注入敏感描述
credentialfeature.aggrega.ai/kind=secret否env是Optional credential the platform presents to the model this connection reaches. When attached, the Recipe delivers it to the platform as OPENAI_API_KEY (a Kubernetes secretKeyRef on helm, the inject pipe on compose); unattached, the platform runs without an injected key.

tools

选择器
component.aggrega.ai/category=ai-tool, tool.aggrega.ai/transport=http
注入
—

经由此套接字的每个连接可填充的槽位

槽位选择器必需注入敏感描述
credentialfeature.aggrega.ai/kind=secret否env是Optional credential the platform presents to the tool this connection reaches, under the request header the tool publishes as credential_header. When attached, the Recipe delivers it as AGGREGA_TOOL_<ALIAS>_CREDENTIAL (a Kubernetes secretKeyRef on helm, the inject pipe on compose), has Langflow store as a global variable the variable holding the header's whole value, and writes that variable's name as the header's value in the MCP Tools node the connection reaches, so the credential is written into no rendered flow. Under any header but Authorization that variable is AGGREGA_TOOL_<ALIAS>_CREDENTIAL itself. Under Authorization, which takes Bearer <credential> as its value, it is AGGREGA_TOOL_<ALIAS>_AUTHORIZATION, holding Bearer and the credential: on helm a reference Kubernetes expands from the delivered Secret when the container starts, on compose the credential as delivered. Langflow stores the variable for the administrator at each sign-in, or, with require_authentication false, once when it creates the administrator, so until then the header carries the variable's name. The build halts where the tool publishes no header to present a credential under.
endpointfeature.aggrega.ai/kind=secret否env是The whole address of the tool this connection reaches, the part carrying its key included, for a server that reads its key from its own address. Langflow reads an MCP server's address only from the flow file it loads, where that key would be written as a literal, so an attachment here halts the build naming the connection, on both targets.

vector_store

选择器
datastore.aggrega.ai/engine in (vector, semantic-memory, search)
注入
env, file

经由此套接字的每个连接可填充的槽位

槽位选择器必需注入敏感描述
credentialfeature.aggrega.ai/kind=secret否env是Optional credential the platform presents to the vector store this connection reaches, such as the API key the store checks. When attached, the Recipe delivers it as AGGREGA_VECTOR_STORE_<ALIAS>_CREDENTIAL (a Kubernetes secretKeyRef on helm, the inject pipe on compose), and, for a store publishing a database and a user, as the password inside AGGREGA_VECTOR_STORE_<ALIAS>_CONNECTION_STRING, written unencoded; unattached, no credential is delivered for the store and that string carries no password.

附加槽位

Feature 附加到此组件时所用的命名绑定点。槽位的选择器限定哪个 Feature 可以填充该槽位,其 inject 列表则指明 Feature 的内容可以如何交付。必需槽位必须在每个实例上完成绑定,可以通过实例上的附加,也可以通过其上层作用域的附加;敏感槽位则声明为用于处理机密内容。Feature 也可以以任何槽位都未声明的名称进行附加。

槽位选择器必需注入敏感描述
admin_passwordfeature.aggrega.ai/kind=secret否env是Password of the administrator account the platform seeds at startup, delivered as LANGFLOW_SUPERUSER_PASSWORD (a Kubernetes secretKeyRef on helm, the inject pipe on compose). Attach it wherever require_authentication is left at its default, since that account is then the only way into the editor; with require_authentication false the platform signs every visitor in and nothing reads this slot.

自注入允许列表

未声明自注入管道。

关系

与其他已发布实体之间的类型化边。没有已发布目标的引用会呈现为断开的边,而不是链接。

使用

被使用

Component 概念参考 →

Aggrega 目录 — 收录组件、功能、配方以及渲染它们的插件的开放索引。

基于共享的 Aggrega 设计系统构建。

隐私