Connect Riverflow MCP to Claude
- Open Claude's custom connector flow.
- Name the connector Riverflow and paste the Claude endpoint above.
- Leave OAuth client credentials empty, connect, then choose your Riverflow team and scopes.
Riverflow MCP
Create images, video and ads with the products, assets and style rules saved in Riverflow—directly from Claude, ChatGPT, Codex, Langdock and other compatible MCP clients.
Connect in three steps
Add Riverflow once, authorize your account and give your AI assistant access to the creative workflows and brand context you choose.
Available on paid plans. Uses the same Riverflow Credits system as the app.
Open Riverflow's guided setup and select Claude, ChatGPT, Codex, Langdock or another compatible client.
Use the client-specific URL below. Riverflow runs as a remote Streamable HTTP MCP server.
Authorize a Riverflow team with OAuth, or add a team-scoped API key when your client uses custom headers. Then choose a brand for your workflow.
https://www.riverflow.ai/api/mcpUse the dedicated URL below when connecting Claude or ChatGPT. Other clients can use the general Streamable HTTP endpoint with a compatible authentication method.
https://www.riverflow.ai/api/mcp/claudehttps://www.riverflow.ai/api/mcp/chatgptNeed more help? Open the guided setup inside Riverflow for the complete client-specific checklist and support options.
Riverflow is a hosted remote MCP server, so there is no local package to install. Client labels can change; use the remote-server or custom-connector flow in your client.
x-api-key, then save.x-api-key.Built for production, not one-off prompts
Your assistant can understand the products, scenes, styles and assets already in Riverflow—then create, edit and organize campaign-ready content without breaking the flow of the conversation.
Stage saved products in curated scenes and generate campaign-ready product photography from a natural-language brief.
Create new images, make precise edits, enhance details and upscale approved assets without rebuilding the work from scratch.
Turn approved creative into product video, generate speech audio with curated Riverflow voices and keep production moving in the same conversation.
Create template-based campaign ads, generate variations and review, approve or retry outputs for the next round of testing.
Let your assistant work with the products, assets, scenes, characters, references and style rules already saved in Riverflow.
Plan and run multi-output creative work, track progress and approve or retry results without duplicating jobs.
Start with a real brief
Riverflow MCP gives your assistant the context and tools to plan the work, create the assets, wait for results and bring the finished output back into the chat.
Launch a photoshoot“Create a full product photoshoot for our berry soda. Use the saved packshot, follow our brand style and return six campaign-ready images.”
Build ad variations“Create three campaign ads from the approved hero image, each with a different copy hook while keeping the product and brand treatment consistent.”
Create motion“Make a short product video from the approved image with slow camera movement, soft shadows and a clean final product hold.”
Plan with brand context“List my products, scenes and style rules, then recommend the strongest setup for a summer launch campaign.”

Account, credits and control
OAuth clients let you choose a Riverflow team and the scopes to grant. Developer clients such as Codex can connect with a team-scoped API key.
Billable MCP actions use the app's Credits system. Depending on your plan, they use subscription, top-up or bonus Credits, or are covered by your team contract.
Your saved products and brand assets remain in your Riverflow workspace, with scoped access and confirmation for protected actions.
Riverflow MCP FAQ
The practical details for using Riverflow's remote MCP server with your AI assistant.
Model Context Protocol (MCP) is an open standard that lets AI assistants use approved external tools and retrieve context. Riverflow's hosted MCP server applies that standard to the products, assets, style rules and creative workflows in your Riverflow team.
Read the Model Context Protocol documentationRiverflow provides guided setup for Claude, ChatGPT, Codex and Langdock. Other clients can connect when they support remote MCP servers over Streamable HTTP and compatible OAuth or custom-header authentication.
You can create product photoshoots, freestyle images, image edits, enhancements, videos, speech audio, ads and batch variations. You can also browse products, assets, scenes, style rules, characters and references before creating. Available tools and result rendering vary by client and the scopes you grant.
Claude, ChatGPT and other supported OAuth clients can connect by signing in to Riverflow, without configuring a client ID or secret. Developer clients that need a custom HTTP header can use a Riverflow API key; the in-app setup explains the correct option for each client.
Billable MCP actions use the same Riverflow Credits system as the app. Depending on your plan, they use subscription, paid top-up or bonus Credits, or are covered by your team contract. Read-only actions do not generate media. Credit use depends on the selected workflow, model and output volume.
Yes. Your assistant can inspect the products, reusable assets, scenes, style rules, characters and references saved for a brand within the authorized Riverflow team, then use that context in supported creative workflows.
No. MCP gives your AI assistant access to Riverflow workflows from the conversation where you already work. The Riverflow app remains the place to manage your brand system, review creative and organize the resulting assets.
For an OAuth connection, reconnect and choose a different team during authorization. If your client uses an API key, replace it with a key from that team. To use another brand in the same team, ask your assistant to list the available brands and select one.
Connect Riverflow to your preferred AI client and turn your brand context into finished product creative.