Voyager Agent runs on your own machine — the client starts it, so there is no endpoint to ping. 141 installs a week from npm. Last commit 21 Jul 2026.
One universal Voyager: name a goal; it activates senses (repo, net) and runs a live mission.
Today is the operative word: we check Voyager Agent every 15 minutes and re-read its code on every release. Watch it and you find out the day that stops being true.
This server runs on your own machine — install it with the package manager and the client starts it for you. Package name taken from the official registry entry.
claude mcp add voyager-agent -- npx -y @dir-ai/voyager-agent
{
"mcpServers": {
"voyager-agent": {
"command": "npx",
"args": [
"-y",
"@dir-ai/voyager-agent"
]
}
}
}
[mcp_servers.voyager-agent]
command = "npx"
args = ["-y", "@dir-ai/voyager-agent"]
{
"mcpServers": {
"voyager-agent": {
"command": "npx",
"args": [
"-y",
"@dir-ai/voyager-agent"
]
}
}
}
{
"servers": {
"voyager-agent": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@dir-ai/voyager-agent"
]
}
}
}
Audits any live MCP server URL or GitHub repo on Glama's 6-dimension TDQS rubric and compiles drop
Local Claude Code and Codex sessions on one machine exchange letters through native inputs.
Read Textline conversations, contacts and reports; send a text, add notes, resolve or transfer.
Publish AI reports as managed native links with versions, access control, revocation, and audit.
Lint OpenAPI 3.x documents for missing security, responses, descriptions and versioning.
One memory for every AI you work with: your rules, notes and decisions in your own Git repo.
Ask a repository what depends on what. Every answer cites the file, line and commit it came from.
AI-native git hosting — repos, PRs, issues, CI gates, and AI code review over MCP (60 tools).
Answers built from our own checks of this server.