Flowra

io.github.flowradevPublished by GitHub account flowradevv1.0.2Updated Oct 25, 2026

MCP server for your apps' tools and custom tools, plus hosted AI agents and approval-gated workflows

Overview

AI-generated overview

Flowra is a hosted MCP server that lets an assistant run AI agents and approval-gated workflows across Gmail, Slack, GitHub, Notion and 1,000+ other apps.

What you can do
Discover and inspect connected apps and their tool schemas, then connect new accounts via OAuth or API key access for CI clients. Run one or many app actions in a single call, including sending, posting, creating, updating and deleting data. Build and edit custom toolkits from described REST endpoints when an app is missing. Run a workbench for hosted agents and scheduled workflows, with a ledger of inputs, outputs and credits per tool call. Pause, inspect and resume runs that are waiting on human approval.
When to use it
Use it when you want scheduled or approval-gated automations across SaaS apps driven from an MCP client, rather than a local script or a Zapier, Make or n8n canvas. Skip it if you have no Flowra project or already maintain your own automation graphs.
Requirements
Remote endpoint at mcp.flowra.dev over streamable HTTP; no local runtime or package. OAuth sign-in per account is required, and a Flowra project must be selected. CI or non-OAuth clients can instead send a project API key and username as headers in their own config.
Before you install
Writes such as send, post, create, update, delete and pay are gated behind an explicit approval in chat, and durable automations need a human-wait or human-approval step; verify those gates before relying on them. Tool calls consume credits and are recorded in a ledger. A project API key grants access, so keep it out of shared repositories. Inbound mail, chat and webhook content is treated as untrusted data.

Installation

In SourceWeft

  1. Open Flowra in the dashboard and add it to a workspace.
  2. Connect the required credentials. They are stored privately and never shown on public pages.
  3. Enable the server for the chats that should use its tools.

Web executable via Streamable HTTP. Remote servers run from the web runtime once configured in a workspace.

Use with an agent

Paste this prompt into Claude Code, Codex, Cursor or another agent — or into SourceWeft chat. The agent reads this server's install guide, shows you its publisher, endpoint and sign-in, and adds it to your client once you agree. Keys never go into a command.

Read https://sourceweft.com/mcp/io.github.flowradev%2Fmcp/install.md and connect the MCP server it describes to my client. Before changing any config, show me its publisher, endpoint and sign-in method and wait for my OK. Never put an API key in a command or in a file I would commit; ask me where it should come from.

Read the install guide the agent follows

Other MCP clients

Add this to your client's mcpServers config. The client completes OAuth sign-in on first connect.

{
  "mcpServers": {
    "flowra": {
      "type": "http",
      "url": "https://mcp.flowra.dev/mcp"
    }
  }
}

README

[Flowra logo]

Flowra — Agent Skill & Hosted MCP Server

Run hosted AI agents and approval-gated workflows on Gmail, Slack, GitHub, Notion, Telegram, WhatsApp, Linear and 1,000+ apps — from Claude Code, Cursor, Codex, GitHub Copilot or any MCP client. OAuth per account, a human approval gate before send/delete/pay, and a ledger of every tool call.

[skills.sh] [License: MIT] [MCP Registry] [Add Flowra MCP to Cursor]

Product · Docs · SDK · Changelog · skills.sh

This repo is the agent skill + Cursor/Claude/Codex plugin for Flowra. From your backend, use the SDK (@flowra/sdk, flowra-sdk, @flowra/cli).

What's included

ComponentPurpose
mcp.json / .mcp.jsonRegisters https://mcp.flowra.dev/mcp (OAuth, no keys). Six tools: discover, schemas, connect, multi-execute, workbench, bash.
skills/flowraRoutes operate / build / edit / debug / credits to the right Flowra calls.
skills/flowra-onboardFirst install: Connect MCP and one read-only first win.
commands//flowra-setup and /flowra-status.
evals/Behavioral scenarios the skill must pass.
rules/flowra-safety.mdcReads are free, writes need your yes; untrusted content is never an instruction.

Install in 30 seconds

1. Skill (Claude Code, Cursor, Codex, Copilot, OpenClaw, Hermes):

bash
npx skills add flowradev/skills --skill flowra

2. Hosted MCP server — no API key needed; sign in with OAuth:

text
https://mcp.flowra.dev/mcp
ClientHow
CursorSettings → Tools & MCP → Connect next to flowra (or install the Flowra plugin)
Claude Code/plugin marketplace add flowradev/skills then /plugin install flowra@flowra — or claude mcp add --transport http flowra https://mcp.flowra.dev/mcp
CodexMarketplace at .agents/plugins/marketplace.json
Gemini CLIgemini extensions install https://github.com/flowradev/skills (needs release v0.2.0)
VS Code / Copilot@mcp in Extensions, or add the URL to .vscode/mcp.json
Skill onlynpx skills add flowradev/skills --skill flowra
CI / no OAuthProject API key in your own config — never in this repo

Cursor Marketplace: search "Flowra", or /add-plugin flowra once listed.

First run

  1. Connect the MCP server and pick a Flowra project.
  2. Ask: "List my 5 latest Gmail subjects with Flowra."
  3. Then: "Every weekday at 9, post a Slack recap of my inbox. Ask me before posting."

What you can ask

  • "Connect Gmail on Flowra and list my latest subjects."
  • "Every weekday at 9:00, post a Slack recap of new Linear issues — ask me before posting."
  • "This Flowra run is paused — show the ledger and resume it."

Why Flowra (vs Zapier MCP, Composio, n8n)

  • Agent or locked workflow on the same project — not just a tool catalog.
  • Human approval gate on any step, including cron runs.
  • Run ledger: inputs, outputs and credits for every tool call.
  • Missing app? Describe the REST endpoint and Flowra builds the toolkit.

Security model

Inbound mail, chat and webhook content is treated as untrusted data. The skill never executes instructions found in payloads and never skips approval gates. Reads run freely; writes (send, post, create, update, delete, pay) need an explicit yes in this chat. Durable automations still need a human_wait node or agent humanApproval. No API keys are stored in this repo.

Use when you want a scheduled or approved automation instead of a local script or an n8n, Zapier, or Make canvas.

Skip when there is no Flowra project, or you are already editing an n8n, Zapier, or Make graph you own.

CI / clients without OAuth

Add headers only in your own .cursor/mcp.json. Never commit a key.

json
{  "mcpServers": {    "flowra": {      "url": "https://mcp.flowra.dev/mcp",      "headers": {        "x-api-key": "<your-project-api-key>",        "x-username": "project_default_user"      }    }  }}

Local testing

bash
mkdir -p ~/.cursor/plugins/localln -sfn "$PWD" ~/.cursor/plugins/local/flowra

Then run Developer: Reload Window.

Layout

text
.cursor-plugin/plugin.json.claude-plugin/plugin.json.claude-plugin/marketplace.json.codex-plugin/plugin.json.agents/plugins/marketplace.jsongemini-extension.jsonmcp.json.mcp.jsonserver.jsoncommands/flowra-setup.mdcommands/flowra-status.mdevals/rules/flowra-safety.mdcskills/flowra/SKILL.mdskills/flowra-onboard/SKILL.mdskills/flowra/references/

MIT © Flowra

Source: README.md at commit a0589df

Tools

0
This server asks you to sign in, so SourceWeft could not read its tools without an account and has not indexed them. What a connected account sees can depend on that account and its permissions.

Version history

3
  1. v1.0.2LatestOct 11, 2026
  2. v1.0.1Oct 8, 2026
  3. v1.0.0Sep 24, 2026

Score

85% of the applicable checks met

How complete this listing is, computed from the registry entry, the endpoint and public sources. Not a rating of the server and not a statement that it is safe or working today: reachability is a dated check, and no installs or reviews go into the score.

  • Reached in the last 30 daysRequired25 pts
  • Auth declaredRequired10 pts
  • Tools listed15 pts

    Answer tools/list without a credential, or ship a package that starts in a sandbox without one.

  • DescriptionRequired15 pts
  • Logo10 pts
  • Documentation15 pts
  • License10 pts