How do I use Tidewave MCP with phxagents?

Install the elixir-phoenix plugin in Claude Code and add Tidewave to your Phoenix app. When the mcp__tidewave__ tools are available in a dev session, phxagents’ tidewave-integration skill activates and routes runtime questions — docs, SQL, schema introspection — through Tidewave instead of shelling out.

Is Tidewave MCP safe to use in production?

No. phxagents treats Tidewave as dev-only. Its tools evaluate code and run SQL against the live runtime, so the plugin never uses them in production contexts and avoids shared dev servers holding copies of production data.

Tidewave MCP for Phoenix

Q: What is Tidewave MCP?

Tidewave MCP is a Model Context Protocol server (from Dashbit) that exposes a running Phoenix application’s runtime to AI editors and coding agents — live SQL queries, code evaluation, Ecto schema and source lookup, logs, and version-exact hex docs. phxagents teaches Claude Code to prefer these runtime tools over guessing from Bash.

Tidewave MCP gives your AI editor runtime intelligence for a running Phoenix app — live SQL, code evaluation, Ecto schema introspection, and version-exact hex docs over the Model Context Protocol. phxagents ships a dedicated Tidewave integration skill so Claude Code prefers these runtime tools over guessing from Bash.

What is Tidewave MCP?

Tidewave (from Dashbit, the team behind Elixir and Livebook) exposes a running Phoenix application’s runtime to AI editors and coding agents through an MCP server. Instead of reading static files and guessing, the agent can query the live database, evaluate code in the running app, look up Ecto schemas and source locations, read logs, and fetch documentation pinned to your mix.lock versions.

phxagents wires this into the Claude Code workflow with the tidewave-integration skill. When the mcp__tidewave__ tools are present in a dev session, the plugin routes runtime questions through Tidewave automatically.

Runtime tools phxagents uses

Each Tidewave MCP tool has a Bash fallback the plugin only reaches for when Tidewave is unavailable:

Task	Tidewave MCP tool	Fallback
Version-exact hex docs	`mcp__tidewave__get_docs`	`web_fetch hexdocs.pm/…`
Evaluate code in the app	`mcp__tidewave__project_eval`	`mix run -e "…"`
Run a SQL query	`mcp__tidewave__execute_sql_query`	`psql $DATABASE_URL`
Find a source location	`mcp__tidewave__get_source_location`	`grep -rn "defmodule"`
List Ecto schemas	`mcp__tidewave__get_ecto_schemas`	Read `lib/*/schemas/`
Read runtime logs	`mcp__tidewave__get_logs`	`tail -f log/dev.log`

Setup & detection

Add Tidewave to your Phoenix app (see tidewave.ai), then confirm the MCP endpoint is reachable in development:

Check the Tidewave MCP endpoint

curl -s http://localhost:4000/tidewave/mcp \
  -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"ping"}'

Inside Claude Code, run /mcp to verify the mcp__tidewave__ tools are connected. Once they are, phxagents’ runtime workflows light up with no extra configuration.

Dev only — never in production

Tidewave tools evaluate code and run SQL against the live runtime, so phxagents enforces a hard rule: Tidewave is dev-only. The plugin never uses these tools in production contexts and avoids shared dev servers that hold copies of production data. SQL is treated as read-heavy — execute_sql_query for SELECTs, with care around mutations.

Learn more

tidewave-integration skill — the full reference phxagents loads on demand
Install phxagents — add 47 skills and 25 agents to Claude Code
SKILL.md source on GitHub
tidewave.ai — the Tidewave runtime intelligence product