CLI Reference: eve init, dev, build, deploy, and More

The eve binary runs from your app root. Every command first loads .env and .env.local from that root before doing anything else. Running eve with no command is equivalent to running eve dev.

Commands at a glance

Command	Description
`eve init [target]`	Scaffold a new agent, or add one to an existing project directory
`eve info`	Print the resolved application: discovered tools, skills, subagents, schedules, channels, routes, artifact paths, and discovery diagnostics
`eve build`	Compile `.eve/` artifacts and build the host output; prints the output directory
`eve start`	Serve the previously built `.output/` app; prints the listening URL
`eve dev`	Start the local dev server and open the terminal UI
`eve dev <url>`	Connect the terminal UI to an existing server URL instead of booting a local server
`eve link`	Link the directory to a Vercel project and pull AI Gateway credentials
`eve deploy`	Deploy the agent to Vercel production (links first if needed)
`eve eval`	Run evals against the local app or a remote target
`eve channels add [kind]`	Scaffold a channel interactively, or by kind (`slack` \| `web`)
`eve channels list`	List user-authored channels in the current project

`eve init`

eve init [target] [--channel-web-nextjs]

The optional target controls the scaffolding mode:

A name (eve init my-agent) scaffolds a fresh project in a new my-agent/ directory, installs dependencies, and initializes Git.
An existing directory (eve init .) adds an agent to that project. Requires a package.json; agent/ files must not exist yet. Adds eve, ai, and zod dependencies without touching anything else.
No target scaffolds or updates the current directory, same as eve init .. Exception: when running inside a coding agent CLI (Claude Code, Cursor, Codex, etc.), a bare eve init prints a setup guide instead of scaffolding.

On an interactive terminal, when a supported coding-agent CLI (claude, codex, cursor-agent, droid, gemini, opencode, or pi) is on PATH, eve init offers available REPLs and eve dev (the default). Each REPL receives a project-specific prompt.

--channel-web-nextjs

flag

Add the Web Chat application (a Next.js app) to the fresh project. Rejected when adding to an existing project — run eve channels add web there instead.

`eve info`

eve info [--json]

Prints the fully resolved application surface: discovered tools, skills, subagents, schedules, channels, routes, artifact paths, and any discovery diagnostics (errors and warnings). Run this first when something behaves unexpectedly. It confirms that a file was discovered and surfaces errors faster than booting the dev server.

--json

flag

Emit output as JSON instead of formatted text.

`eve build`

eve build

Compiles .eve/ artifacts and builds the host output, then prints the built output path. No flags. When eve build fails on discovery errors, it prints the full diagnostics report (severity, message, source path) and the diagnostics artifact path.

Artifacts written under `.eve/`

These artifacts are preserved even on partial failure and are useful for debugging:

Artifact	Description
`.eve/discovery/agent-discovery-manifest.json`	What eve found on disk
`.eve/discovery/diagnostics.json`	Authored-shape errors and warnings
`.eve/compile/compiled-agent-manifest.json`	The serialized authored surface eve loads at runtime
`.eve/compile/compile-metadata.json`	Build-time metadata and paths
`.eve/compile/module-map.mjs`	Compiled module entrypoints eve imports at runtime

`eve start`

eve start [--host <host>] [--port <port>]

Serves the previously built output. Prints the listening URL.

--host

string

Host interface to bind. Defaults to all interfaces.

--port

number

Port to listen on. Defaults to $PORT, then 3000.

`eve dev`

eve dev [options]
eve dev https://your-app.vercel.app

Starts the local dev server with hot module replacement and opens the interactive terminal UI. Pass a bare URL as the only argument to connect the UI to an existing server instead of booting a local one — useful for smoke-testing a preview or production deployment. The interactive UI turns off automatically in a non-TTY terminal.

--host

string

Host interface to bind. Defaults to all interfaces.

--port

number

Port to listen on. Defaults to $PORT, then 3000.

-u, --url

string

Connect to an existing server URL instead of starting one.

--no-ui

flag

Start the server without an interactive terminal UI. Use this for headless verification and CI.

--name

string

Title shown in the terminal UI. Defaults to the app folder name.

--input

string

Pre-fill the prompt input after launching the UI (editable, not auto-submitted).

--tools

enum

Tool-call rendering mode: full | collapsed | auto-collapsed | hidden. Default: auto-collapsed.

--reasoning

enum

Reasoning rendering mode: full | collapsed | auto-collapsed | hidden. Default: full.

--subagents

enum

Subagent-section rendering mode: full | collapsed | auto-collapsed | hidden. Default: auto-collapsed.

--connection-auth

enum

Connection-authorization rendering mode: full | collapsed | auto-collapsed | hidden. Default: full.

--assistant-response-stats

enum

Assistant header statistic: tokens | tokensPerSecond. Default: tokensPerSecond.

--context-size

number

Model context window size in tokens, shown as a usage percentage in the UI.

--logs

enum

Server/agent log streams to show: all | stderr | sandbox | none. Default: stderr.

Local dev writes the active server process ID to .eve/dev-process.pid. If another eve dev starts for the same agent while that process is still running, eve exits with a message that includes the command to stop the existing server.

`eve link`

eve link

Links the current directory to an existing Vercel project and pulls AI Gateway credentials into .env.local. You select a team and then one of its projects; type a name and choose Search for ‘<name>’ to search further. Running eve link again re-links: the pickers always run and the new choice wins. This command is interactive only — in CI, use vercel link --project <name> --yes --non-interactive instead. A running eve dev reloads env files automatically after linking; no restart is needed.

`eve deploy`

eve deploy

Deploys the agent to Vercel production (vercel deploy --prod), installing dependencies first and pulling environment variables after. An already-linked project deploys with or without a TTY. An unlinked directory walks the eve link pickers when a terminal is present, and exits with guidance otherwise.

`eve eval`

eve eval [evalId...] [--url <url>] [options]

Runs all discovered evals when no IDs are given. IDs match exactly or by directory prefix — eve eval weather runs everything under evals/weather/. Exit codes:

0 — every eval passed its checks
1 — any eval failed (failed check, execution error, or --strict threshold miss)
2 — configuration errors

--url

string

Remote agent URL to target. Skips local host startup.

--tag

string

Run only evals carrying a specific tag. Repeatable.

--strict

flag

Below-threshold scores also fail the exit code.

--list

flag

Print discovered evals without running them.

--timeout

number

Per-eval timeout in milliseconds.

--max-concurrency

number

Maximum number of concurrent eval executions. Default: 8.

--json

flag

Output results as JSON.

--junit

string

Write JUnit XML results to the specified file path.

--skip-report

flag

Skip eval-defined reporters (e.g. Braintrust).

--verbose

flag

Stream per-eval t.log lines to stdout.

`eve channels add`

eve channels add [kind] [-f] [-y]

Scaffolds a channel into agent/channels/. With no kind argument, prompts interactively. Pass a kind to scaffold directly: slack or web.

kind

string

Channel kind to scaffold: slack | web. Omit to choose interactively.

-f, --force

flag

Overwrite existing channel files.

-y, --yes

flag

Assume yes for confirmations. Requires an explicit kind.

`eve channels list`

eve channels list [--json]

Lists the user-authored channels in the current project.

--json

flag

Output as JSON.

Recommended development loop

# 1. Edit files under agent/
# 2. Confirm discovery
eve info

# 3. Iterate locally
eve dev

# 4. Prepare for production
eve build

# 5. Smoke-test the build
eve start

Use eve dev --no-ui for headless verification in automated environments or when a coding agent is running the server.

Project Layout

What eve info discovers on disk

TypeScript API

define* helpers and import paths

API Reference

Concepts

CLI Reference: eve init, dev, build, deploy, and More

Commands at a glance

`eve init`

`eve info`

`eve build`

Artifacts written under `.eve/`

`eve start`

`eve dev`

`eve link`

`eve deploy`

`eve eval`

`eve channels add`

`eve channels list`

Recommended development loop

Project Layout

TypeScript API

Build docs developers (and LLMs) love

API Reference

Concepts

Documentation Index

​Commands at a glance

​eve init

​eve info

​eve build

​Artifacts written under .eve/

​eve start

​eve dev

​eve link

​eve deploy

​eve eval

​eve channels add

​eve channels list

​Recommended development loop

Project Layout

TypeScript API

Build docs developers (and LLMs) love

Commands at a glance

`eve init`

`eve info`

`eve build`

Artifacts written under `.eve/`

`eve start`

`eve dev`

`eve link`

`eve deploy`

`eve eval`

`eve channels add`

`eve channels list`

Recommended development loop