devto-mcp
A work management MCP server for Claude Code that integrates with Jira Cloud to handle sprints, tickets, and project tracking via natural language.
README
devto-mcp
The work management layer for Claude Code.
DevTo lives inside Claude Code as an MCP server. It manages your sprint, your tickets, and your project tracker — without you ever leaving your terminal. Status updates, ambient comments on transitions, plan generation from natural language, semantic ticket matching. All from your editor.
Today: full Jira Cloud support via one-click OAuth. Linear and GitHub Issues coming.
Install
npx devto-mcp@latest --help
You don't need a global install. devto init writes a project-scoped .mcp.json that runs the latest published version via npx on each Claude Code session start.
Requires Node.js 18+ and Claude Code.
Setup
1. Create an account
Sign up at devto.ai. Connect your Jira workspace from the onboarding screen — click Connect with Jira and approve the OAuth consent. No tokens to copy, no API keys to paste.
2. Link the project
In your project directory:
cd your-project
npx devto-mcp@latest devto init
A short verification code appears. The CLI opens your browser to devto.ai/cli/authorize — confirm the code matches, click Approve, and the CLI receives your API key silently.
devto init writes:
| File | Purpose |
|---|---|
.devto.json |
Project metadata (workspace, project key, provider) |
.devto/config.json |
Local credentials (API key, scoped to this project) |
.mcp.json |
Claude Code MCP server config (npx command + env) |
.devto/, .devto.json, and .mcp.json are added to .gitignore automatically.
3. Restart Claude Code
MCP server configs load at session start. Quit Claude Code fully and reopen — claude mcp list should now show devto.
You're done. No Anthropic key required: AI planning runs through DevTo's backend at no per-call cost to you (subject to your plan's monthly action quota).
Use it
Just talk to Claude Code about your work. The MCP guides Claude toward the right tools — you don't need to memorize them.
> what's in my current sprint?
> create a plan for adding Stripe webhooks
> mark the auth task as done and add a comment about the redirect fix
> show recent comments on JOB-22
DevTo's natural-language semantic matching means you can refer to tickets by description ("the login bug", "that auth task") instead of memorizing keys like JOB-22. Ambiguous matches surface a disambiguation prompt.
Tools
44 MCP tools, grouped by area. Claude Code reads the full descriptions from the server at session start.
Project overview
get_project_summary— lightweight snapshot of active work + epic progress (called at session start)get_status— aggregate project metricslist_epics— epic progress barsget_epic— deep dive into a single epic
Planning
create_plan— AI-generated epic + stories + subtasks (preview only)confirm_plan— execute a previously-previewed plan
Issue management
create_epic,create_task,create_subtask— direct createsupdate_task,update_epic— status, fields, comments. Accepts natural language refs.delete_issue— destructive, asks for confirmationget_tasks— list open tasks, optionally filter by epic
Comments
add_comment,get_comments,delete_comment
Sprint management
list_sprints,create_sprint,move_to_sprint,manage_sprint
Search
search_issues— full JQL support
Field discovery
get_priorities,get_statuses,get_issue_types,get_labels,get_components,get_assignable_users,get_custom_fields
Issue linking
link_issues,get_issue_links,get_link_types
Time tracking
log_work,get_worklog,set_estimate
Versions / releases
list_versions,create_version,release_version
Backlog
get_backlog,move_to_backlog
Bulk operations
bulk_create_tasks,bulk_update_tasks
Attachments
attach_file,list_attachments
Watchers
manage_watchers
CLI commands
| Command | What it does |
|---|---|
devto login |
Authenticate globally (browser flow) |
devto init |
Link the current directory to a DevTo workspace |
devto status |
Show connection state + active project |
devto doctor |
Diagnostic — checks API key, workspace token, MCP registration, AI mode |
devto sync |
Re-fetch workspace config from devto.ai |
devto unlink |
Remove this project's link (keeps the package installed) |
devto config set anthropic-key <key> |
Optional Expert Mode — see below |
Expert Mode (BYOK)
By default, DevTo's backend handles Anthropic calls for create_plan. If you'd rather use your own Anthropic key — for billing reasons, or because you have credits to burn — run:
devto config set anthropic-key sk-ant-xxxx
DevTo will detect the local key and call Anthropic directly from your machine instead of the server proxy. Your code context still never leaves your machine in either mode; the only thing that changes is who pays for the inference.
Per-project configuration
Each project gets its own .devto.json:
{
"workspaceUrl": "https://yourcompany.atlassian.net",
"projectKey": "ENG",
"provider": "jira"
}
The MCP server walks up the directory tree from cwd looking for .devto.json. Found? That workspace + project apply. Not found? Falls back to the tenant's active workspace from the dashboard.
You can have many projects with different .devto.json files on the same machine. Each one is independent.
Troubleshooting
Claude Code doesn't see devto's tools
- Did you restart Claude Code after
devto init? MCP configs load at session start, not live. - Run
claude mcp listfrom the project directory. You should seedevto:in the output. - Try
devto doctor— it prints a complete diagnostic with green checkmarks or red Xs and a suggested fix for each.
OAUTH_REAUTH_REQUIRED error mid-session
Your Jira connection's OAuth token has expired or been revoked. Visit devto.ai/dashboard/settings/workspace and click Reconnect. Takes 10 seconds.
Need to start over for a project
devto unlink clears .devto.json, .devto/, the devto entry in .mcp.json, and the project's row in ~/.devto/projects.json. Other projects are unaffected.
Dashboard
Manage workspaces, API key sessions, billing, and team at devto.ai/dashboard.
License
MIT — see LICENSE.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
Neon Database
MCP server for interacting with Neon Management API and databases
E2B
Using MCP to run code via e2b.
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.