What can you do with this server?

RelayPlane is an MCP server that reduces AI context usage by 90%+ in multi-step workflows by keeping intermediate results in the workflow engine instead of passing them through your context window. Core Capabilities: * Execute single AI prompts - Test individual model calls across multiple providers (OpenAI, Anthropic, Google, xAI) with structured output support and cost tracking * Run multi-step AI workflows - Orchestrate complex pipelines where steps reference previous outputs using template variables ({{input.field}}, {{steps.stepName.output}}), dramatically reducing context usage * Validate workflow structure - Check DAG structure, dependencies, and model formats before execution without making API calls (free) * Browse available AI models - View models from OpenAI (GPT-5.2, o3-mini), Anthropic (Claude Opus/Sonnet/Haiku 4.5), Google (Gemini 3), and xAI (Grok) with capabilities and pricing * Access pre-built workflow templates - Use production-ready patterns for invoice processing (97% context reduction), content pipelines (90%), and lead enrichment (80%) * Track execution history - Review recent runs with full details including step outputs, token usage, costs, and trace URLs for debugging * Integrate MCP tools - Call other MCP server tools within workflows using server:tool format * Budget protection - Configure daily spending limits ($5 default), per-call caps ($0.50 default), and rate limiting (100 calls/hour default) * Structured outputs - Define JSON schemas to ensure consistent, validated data formats Key Benefits: BYOK (Bring Your Own Keys) with no markup or RelayPlane fees, multi-provider support in the same workflow, and massive token savings by keeping intermediate results server-side.

Which integrations are available for this server?

Enables AI workflow orchestration using Google's AI models for single prompts and multi-step workflows with context reduction. Enables AI workflow orchestration using OpenAI models (e.g., gpt-4o) for single prompts and multi-step workflows with context reduction.

How do I use RelayPlane?

1. Click on "Install Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@RelayPlane create a workflow to summarize this article and generate social media posts from it" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

RelayPlane MCP Server

Reduce AI context usage by 90%+ in multi-step workflows

RelayPlane keeps intermediate results in the workflow engine instead of passing them through your context window—saving tokens and reducing costs.

Quick Start

1. Install with API Keys (Recommended)

claude mcp add relayplane \ -e OPENAI_API_KEY=sk-... \ -e ANTHROPIC_API_KEY=sk-ant-... \ -- npx @relayplane/mcp-server

2. Restart Claude Code

Important: You must fully restart Claude Code after adding the MCP server. The /mcp command only reconnects—it doesn't reload environment variables.

3. Test the Connection

Ask Claude: "Use relay_models_list to show configured providers"

Models should show configured: true for providers with valid API keys.

Installation Options

Option A: Inline API Keys (Simplest)

claude mcp add relayplane \ -e OPENAI_API_KEY=sk-proj-... \ -e ANTHROPIC_API_KEY=sk-ant-... \ -e GOOGLE_API_KEY=AIza... \ -e XAI_API_KEY=xai-... \ -- npx @relayplane/mcp-server

Option B: Shell Environment Variables

First, add to your shell profile (~/.zshrc or ~/.bashrc):

export OPENAI_API_KEY=sk-proj-... export ANTHROPIC_API_KEY=sk-ant-... export GOOGLE_API_KEY=AIza... export XAI_API_KEY=xai-...

Then source and install:

source ~/.zshrc claude mcp add relayplane -- npx @relayplane/mcp-server

Option C: Manual Configuration

Edit ~/.claude.json directly:

{ "projects": { "/your/project/path": { "mcpServers": { "relayplane": { "type": "stdio", "command": "npx", "args": ["@relayplane/mcp-server"], "env": { "OPENAI_API_KEY": "sk-proj-...", "ANTHROPIC_API_KEY": "sk-ant-...", "GOOGLE_API_KEY": "AIza...", "XAI_API_KEY": "xai-..." } } } } } }

Warning: The env field must contain actual API keys, not variable references like ${OPENAI_API_KEY}. Variable substitution is not supported in the MCP config file.

Model IDs

Important: Always check https://relayplane.com/docs/providers for the latest model IDs. The relay_models_list tool may return outdated information.

OpenAI — prefix: `openai:`

Model ID	Best For
gpt-5.2	Latest flagship, 1M context
gpt-5-mini	Cost-efficient, fast
gpt-5-nano	Ultra-fast, minimal cost
o3-mini	Reasoning tasks

Anthropic — prefix: `anthropic:`

Model ID	Best For
claude-opus-4-5-20251101	Most intelligent, complex tasks
claude-sonnet-4-5-20250929	Best coding, strongest for agents
claude-haiku-4-5-20251001	Fast, high-volume tasks
claude-3-5-haiku-20241022	Fast, affordable (legacy)

Google — prefix: `google:`

Model ID	Best For
gemini-3-pro	Most powerful multimodal
gemini-3-flash	Fast multimodal
gemini-2.5-flash	Cost-effective

xAI — prefix: `xai:`

Model ID	Best For
grok-beta	Latest flagship, 256K context

Example Usage

{ "name": "my-step", "model": "openai:gpt-5.2", "prompt": "Analyze this data..." }

Available Tools

Tool	Purpose	Cost
relay_run	Single prompt execution	Per-token
relay_workflow_run	Multi-step orchestration	Per-token
relay_workflow_validate	Validate DAG structure	Free
relay_skills_list	List pre-built patterns	Free
relay_models_list	List available models	Free
relay_runs_list	View recent runs	Free
relay_run_get	Get run details	Free

Budget Protection

Default safeguards (customizable via CLI flags):

Limit	Default	Flag
Daily spending	$5.00	--max-daily-cost
Per-call cost	$0.50	--max-single-call-cost
Hourly requests	100	--max-calls-per-hour

RelayPlane is BYOK (Bring Your Own Keys)—we don't charge for API usage. Costs reflect only your provider bills.

Pre-built Skills

Use relay_skills_list to see available workflow templates:

Skill	Context Reduction	Use Case
invoice-processor	97%	Extract, validate, summarize invoices
content-pipeline	90%	Generate and refine content
lead-enrichment	80%	Enrich contact data

Configuration

Persistent Config File

Create ~/.relayplane/mcp-config.json:

{ "codegenOutDir": "./servers/relayplane", "maxDailyCostUsd": 10.00, "maxSingleCallCostUsd": 1.00, "maxCallsPerHour": 200 }

Note: API keys should be passed via environment variables or the Claude Code MCP env field—not stored in this config file.

Troubleshooting

"Provider not configured" Error

Provider "openai" (step "extract") is not configured. Set OPENAI_API_KEY environment variable.

Causes:

API key not passed to MCP server
Claude Code not restarted after config change

Solutions:

Check your MCP config in ~/.claude.json:

"relayplane": { "env": { "OPENAI_API_KEY": "sk-..." // Must be actual key, not ${VAR} } }

Fully restart Claude Code (exit with Ctrl+C, relaunch)
Verify configuration: Ask Claude: "Use relay_models_list and check which show configured: true"

Model Not Found (404 Error)

Anthropic API error: 404 - model: claude-3-5-sonnet-20241022

Cause: Model ID is outdated or incorrect.

Solution: Check current model IDs at: https://relayplane.com/docs/providers

Common fixes:

Use claude-sonnet-4-5-20250929 for latest Claude Sonnet
Use gpt-5.2 for latest OpenAI flagship model

Config Changes Not Taking Effect

Cause: /mcp reconnect doesn't reload environment variables.

Solution: Fully restart Claude Code:

Exit with Ctrl+C
Relaunch claude
Run /mcp to verify connection

Workflow Validation Passes But Execution Fails

Cause: relay_workflow_validate only checks DAG structure, not:

API key validity
Model availability
Schema compatibility

Solution: Test with a simple relay_run first:

Use relay_run with model "openai:gpt-5.2" and prompt "Say hello"

Quick Test

After setup, verify everything works:

Use relay_workflow_run to create an invoice processor: - Step 1 (extract): Use openai:gpt-5.2 to extract vendor, total from invoice - Step 2 (validate): Use anthropic:claude-haiku-4-5-20251001 to verify math Input: "Invoice from Acme Corp, Total: $500"

Expected: Both steps complete successfully with structured output.

Support

Documentation: https://relayplane.com/docs
Model IDs: https://relayplane.com/docs/providers
Issues: https://github.com/RelayPlane/mcp-server/issues

License

MIT

Install Server

A

security – no known vulnerabilities

F

license - not found

A

quality - confirmed to work

How are these scores calculated?

Resources

Need Help?

Report Issue

Related Servers

RelayPlane

RelayPlane MCP Server

Table of Contents

Quick Start

1. Install with API Keys (Recommended)

2. Restart Claude Code

3. Test the Connection

Installation Options

Option A: Inline API Keys (Simplest)

Option B: Shell Environment Variables

Option C: Manual Configuration

Model IDs

OpenAI — prefix: `openai:`

Anthropic — prefix: `anthropic:`

Google — prefix: `google:`

xAI — prefix: `xai:`

Example Usage

Available Tools

Budget Protection

Pre-built Skills

Configuration

Persistent Config File

Troubleshooting

"Provider not configured" Error

Model Not Found (404 Error)

Config Changes Not Taking Effect

Workflow Validation Passes But Execution Fails

Quick Test

Support

License

Resources

Tools

New MCP Servers

Latest Blog Posts

MCP directory API

RelayPlane MCP Server

Table of Contents

Quick Start

1. Install with API Keys (Recommended)

2. Restart Claude Code

3. Test the Connection

Installation Options

Option A: Inline API Keys (Simplest)

Option B: Shell Environment Variables

Option C: Manual Configuration

Model IDs

OpenAI — prefix: openai:

Anthropic — prefix: anthropic:

Google — prefix: google:

xAI — prefix: xai:

Example Usage

Available Tools

Budget Protection

Pre-built Skills

Configuration

Persistent Config File

Troubleshooting

"Provider not configured" Error

Model Not Found (404 Error)

Config Changes Not Taking Effect

Workflow Validation Passes But Execution Fails

Quick Test

Support

License

Resources

Tools

New MCP Servers

Latest Blog Posts

MCP directory API

OpenAI — prefix: `openai:`

Anthropic — prefix: `anthropic:`

Google — prefix: `google:`

xAI — prefix: `xai:`