AgentDrive MCP Server is a shared workspace backend for coding agents — it stores versioned markdown files, provides semantic memory/search across your org, and tracks provenance — with 20 tools exposed in this schema (README advertises 22; drive_sessions/drive_resume/drive_connect are not in the schema).
Files: drive_write (create/update with auto-versioning), drive_read, drive_list (path-prefix filter), drive_delete, drive_versions (history).
Memory & search: drive_search (org-wide semantic search with agent/path filters, top-k up to 50), drive_recall (persistent facts/preferences), drive_context (auto-pull relevant prior artifacts for a task).
Collaboration: drive_share (view/comment/edit links, optional email gating, expiry, max uses), drive_comment (with line ranges), drive_feedback (read human comments).
Provenance: drive_activity (recent writes/comments/shares), drive_log_run (log task, status, artifacts referenced/produced, duration, metadata).
Org management: drive_org (agents, members, plan, usage), drive_invite (email invite with owner/admin/member role), drive_roles (assign/list/remove per-workspace view/comment/edit access).
Sandbox (experimental): drive_sandbox_clone (git clone into per-workspace Linux sandbox), drive_sandbox_exec (shell commands, 64 KB output cap, 60 s max), drive_sandbox_read, drive_sandbox_commit — all require Cloudflare Containers and degrade to 503 otherwise.
Access: any tool works from a workspace ID you have rights to; connect via hosted Streamable HTTP or local stdio with an API key.
Allows CrewAI agents to operate within a shared workspace with persistent storage, version history, and provenance attribution across sessions.
Enables LangGraph agents to share a persistent workspace with file versioning, semantic recall, and run logging, maintaining context across sessions and tools.
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@AgentDrive MCP Server load context from my last session"
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.
The official Model Context Protocol server for AgentDrive: shared workspace for coding agents. Capture sessions from Claude Code, Codex, OpenCode, Cursor, Gemini CLI and Antigravity, resume them in any of the others, and search everything they wrote. Provenance on every artifact; portable storage on R2, S3 or your own bucket.
Hit a usage limit in one agent? Call drive_resume in another and it loads the
handoff: the goal, the files touched and the latest turns. 22 MCP tools (18
stable + 4 experimental sandbox), over stdio or hosted. Free plan, no credit card.
Website: https://agentdrive.sh
npm: @adrive/mcp-server
Hosted endpoint: https://mcp.agentdrive.sh/mcp
This repo is also the Claude Code plugin marketplace. The plugin captures every session in the background and resumes it from Codex, OpenCode, Gemini CLI, Antigravity, or Cursor:
/plugin install agentdrive --marketplace agentdrive-sh/agentdrive-mcpOn Claude Code older than 2.1.275, run /plugin marketplace add agentdrive-sh/agentdrive-mcp
then /plugin install agentdrive@agentdrive. Sign in with /agentdrive:login.
Related MCP server: Coding Tools MCP
Create one at https://app.agentdrive.sh — no credit card. The free tier (1 workspace, 1 GB storage, 10k embeddings) is free indefinitely.
Point any MCP client that supports remote servers at the hosted endpoint and pass your key as a bearer token:
URL: https://mcp.agentdrive.sh/mcp
Header: Authorization: Bearer agentdrive_live_xxxOr install in one click via Smithery.
{
"mcpServers": {
"agentdrive": {
"command": "npx",
"args": ["-y", "@adrive/mcp-server"],
"env": {
"AGENTDRIVE_API_KEY": "agentdrive_live_xxx"
}
}
}
}Add that block to your host's MCP config:
Claude Desktop: claude_desktop_config.json
Cursor: ~/.cursor/mcp.json
Claude Code: install the plugin (see above), which also captures sessions
Codex / Cline / Continue / Zed: the standard mcpServers block above
The fastest path is the AgentDrive installer, which auto-wires every detected host:
curl -fsSL agentdrive.sh/install | shEnv var | Required | Default | Description |
| no | key saved by | Your AgentDrive API key ( |
| no |
| Override the API endpoint (self-hosted worker). |
| no |
| Which agent runtime is calling — drives cross-agent provenance. One of |
| no | — | Free-text role label (e.g. |
22 tools across files, memory, sessions, collaboration, org, and provenance. The stdio server derives the current project from its working directory (git remote, else repo path), which scopes drive_sessions and drive_resume.
Tool | Category | Description |
| memory | Auto-pull the most relevant prior artifacts for the task you're starting. |
| memory | Semantic search across your org's artifacts. |
| memory | Recall persistent facts/preferences agents have written. |
| sessions | List recent agent sessions for this project across Claude Code, Codex, OpenCode, Gemini CLI, Antigravity, and Cursor. |
| sessions | Load a compact handoff (goal, files touched, recent turns) from a prior session, by default the latest from a different harness. |
| files | Write a markdown file (auto-versioned). |
| files | Read a file's content. |
| files | List files in a workspace. |
| files | Delete a file. |
| files | View a file's version history. |
| collaboration | Generate a (optionally email-gated) share link. |
| collaboration | Comment on a file. |
| collaboration | Read human comments on a file. |
| provenance | Recent workspace activity (writes, comments, shares). |
| provenance | Log an agent run (task, status, artifacts, identity). |
| org | Invite a human to your org by email. |
| org | Assign/list/remove per-workspace access by email. |
| org | View org details: agents, members, plan, usage. |
| sandbox (experimental) | Clone a git repo into a per-workspace Linux sandbox. |
| sandbox (experimental) | Run a shell command in the sandbox. |
| sandbox (experimental) | Read a file from the sandbox filesystem. |
| sandbox (experimental) | Stage + commit the sandbox worktree. |
The drive_sandbox_* tools require the Cloudflare Containers plan on the AgentDrive
worker and degrade cleanly to a "not configured" message otherwise.
Agents lose context at every session boundary and every tool switch. AgentDrive is the shared substrate: one workspace, every agent, full provenance. Write once with Claude Code; recall it from Cursor; hand off to Codex — each call is attributed to the right principal and platform.
npm install
npm run build # tsc -> dist/
npm test # vitestApache-2.0. See LICENSE.
Get recent activity in a workspace (file writes, comments, shares).
Inputs: workspace_id (required).
Add a comment to a file.
Inputs: body (required), file_id (required), workspace_id (required), line_end, +1 more.
Auto-pull the most relevant artifacts from prior agent sessions for the task you are about to do. Combines semantic relevance with recency. Call this at the START of any non-trivial task to ground the agent in prior context — replaces manual copy-paste of artifacts between sessions.
Inputs: task (required), limit, workspace_id.
Delete a file from a workspace.
Inputs: file_id (required), workspace_id (required).
Read human comments on a file. Use this to check if humans have left feedback.
Inputs: file_id (required), workspace_id (required).
Invite a human to your organization by email. They will receive a verification code to join.
Inputs: email (required), role.
List files in a workspace, optionally filtered by path prefix.
Inputs: workspace_id (required), prefix.
Log an agent run to AgentDrive's provenance store. Captures task, status, referenced/produced artifacts, and cross-platform identity (which agent on which tool). Used for retrospective context retrieval and team-wide audit.
Inputs: task (required), run_id, status, metadata, +3 more.
View your organization details: agents, members, plan, and usage.
Inputs: no named parameters.
Read a file's content from a workspace.
Inputs: file_id (required), workspace_id (required).
Recall persistent facts and preferences your agents have written. In v1 this returns semantic search results across artifacts; v1.5 will narrow to auto-extracted facts (preferences, decisions, key entities).
Inputs: query (required), top_k.
Manage per-workspace role delegation. Assign, list, or remove access for specific emails (view/comment/edit).
Inputs: action (required), workspace_id (required), email, permission.
[EXPERIMENTAL] Clone a git repo into a per-workspace Linux sandbox. Requires Cloudflare Containers on the AgentDrive worker; gracefully degrades to a 503 with "sandbox not configured" otherwise.
Inputs: url (required), workspace_id (required), depth, branch.
[EXPERIMENTAL] Stage + commit the current sandbox worktree. Surfaces as an AgentDrive artifact at commits/.json. Requires Cloudflare Containers; degrades to 503 otherwise.
Inputs: message (required), workspace_id (required), author_name, author_email.
[EXPERIMENTAL] Run a shell command inside the per-workspace Linux sandbox. Captures stdout/stderr (64 KB cap), exit code, duration. Requires Cloudflare Containers; degrades to 503 otherwise.
Inputs: cmd (required), workspace_id (required), cwd, timeout_ms.
[EXPERIMENTAL] Read a file from the per-workspace sandbox filesystem (build artifacts, test output, logs). Requires Cloudflare Containers; degrades to 503 otherwise.
Inputs: path (required), workspace_id (required).
Semantic search across your org's artifacts in AgentDrive. Returns the top-N most relevant files. Use this BEFORE writing similar artifacts to avoid duplication and surface prior context.
Inputs: query (required), top_k, filter, workspace_id.
View version history for a file in a workspace.
Inputs: file_id (required), workspace_id (required).
Write a markdown file to a workspace. Creates or updates the file with automatic versioning.
Inputs: path (required), content (required), workspace_id (required), message.
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 20 tools
Tools are generally distinct, with clear verb-noun names. Some overlap exists between drive_search, drive_recall, and drive_context, as all retrieve information; however, descriptions clarify their intended use cases.
All tools follow a consistent 'drive_' prefix with snake_case verb-noun or noun-verb pattern. No mixed conventions or irregular naming.
20 tools cover a broad range of workspace management, collaboration, and experimental sandbox features. The count is slightly high but justified by the diverse functionality.
Core file operations (CRUD, sharing, commenting, versioning) are covered, along with user management, search, and context retrieval. Missing explicit workspace creation/renaming, but write and list imply workspace existence.
Persistent docs and memory for AI agents — read, write, organize & search a shared workspace.
Shared project context for AI agents and teams: docs, tasks, and messages that stay current.
Securely search and manage workspace context files for AI agents and teams.
Shared workspace for AI artifacts: your agents publish, search and cite each other's work.
