Using Pipeboard with OpenAI Codex
Connect Meta Ads and Google Ads to Codex via Streamable HTTP MCP servers. Manage campaigns, analyze performance, and optimize spend — all from Codex.
Before You Start
You need two things: a Pipeboard account and an API token.
Pipeboard Account
Free tier availableSign up at Pipeboard and connect your Meta Ads or Google Ads account.
Create accountAPI Token
For authenticationGenerate an API token from your Pipeboard dashboard. Codex will send it as a Bearer token.
Get your tokenQuick Setup with the CLI
Use codex mcp add to register Pipeboard MCP servers in seconds.
Set Your API Token
Export your Pipeboard API token as an environment variable. Add this to your shell profile (e.g. ~/.zshrc) so it persists.
export PIPEBOARD_API_TOKEN="your-token-here"Add Meta Ads MCP Server
Register the Meta Ads MCP server as a Streamable HTTP endpoint
codex mcp add meta-ads \
--type http \
--url https://meta-ads.mcp.pipeboard.co/ \
--bearer-token-env PIPEBOARD_API_TOKENAdd Google Ads MCP Server
Register the Google Ads MCP server (same token works for both)
codex mcp add google-ads \
--type http \
--url https://google-ads.mcp.pipeboard.co/ \
--bearer-token-env PIPEBOARD_API_TOKENVerify and Use
Check your MCP servers are registered, then start using Codex
# Check registered MCP servers
codex mcp
# In the Codex TUI, use /mcp to see active servers
# Then ask Codex:
> Show me my Meta Ads campaigns from last week
> What are my top performing Google Ads campaigns?Codex will automatically discover all available tools from each MCP server.
Setup via config.toml
Edit ~/.codex/config.toml directly for full control over server configuration.
Add to ~/.codex/config.toml
Add both MCP servers as Streamable HTTP endpoints with bearer token authentication. In the Codex IDE extension, you can open this file from the gear menu under MCP settings > Open config.toml.
[mcp_servers.meta-ads]
url = "https://meta-ads.mcp.pipeboard.co/"
bearer_token_env_var = "PIPEBOARD_API_TOKEN"
[mcp_servers.google-ads]
url = "https://google-ads.mcp.pipeboard.co/"
bearer_token_env_var = "PIPEBOARD_API_TOKEN"Set the Environment Variable
Add your API token to your shell profile so Codex can read it at startup.
export PIPEBOARD_API_TOKEN="your-token-here"Get your token from pipeboard.co/api-tokens. Then reload your shell with source ~/.zshrc.
Optional: Advanced Configuration
You can tune timeouts or restrict which tools Codex can access.
[mcp_servers.meta-ads]
url = "https://meta-ads.mcp.pipeboard.co/"
bearer_token_env_var = "PIPEBOARD_API_TOKEN"
tool_timeout_sec = 120 # allow longer API calls
startup_timeout_sec = 15
# enabled_tools = ["get_campaigns", "get_insights"]
# disabled_tools = ["update_campaign"]Use enabled_tools to allow only specific tools, or disabled_tools to block specific ones. Useful for read-only setups.
Tips for Codex
Get the most out of Pipeboard with Codex
Works in CLI and IDE extension:
The config.toml is shared between the Codex CLI and the IDE extension. Configure once, use everywhere.
Use /mcp in the TUI to verify:
Inside the Codex terminal UI, type /mcp to see which MCP servers are active and their available tools.
Project-scoped config for teams:
Create a .codex/config.toml in your project root for team-specific MCP setups. This only works in trusted projects.
Restrict tools for safety:
Use enabled_tools or disabled_tools in config.toml to limit which ad operations Codex can perform. Great for read-only analysis setups.
What You Can Do
Audit all campaigns and create a performance report
Find underperforming ads and pause them
Duplicate a winning ad with different creative
Create a new campaign from scratch with targeting
Get insights across Meta and Google in one conversation
Bulk update budgets across multiple ad sets
Need Help?
We can help you get set up quickly
Related Guides
Continue learning with these recommended guides.