StateLode
Share a single task layer across AI coding agents with task_search, task_get, and task_mutate instead of scattered todos in chat.
Overview
StateLode is an MCP server for the Build phase that exposes a shared task layer for AI coding agents via task_search, task_get, and task_mutate.
What is this MCP server?
- Unified MCP surface: task_search, task_get, task_mutate
- Hosted API at api.statelode.dev/mcp with Bearer workspace token
- Optional stdio bridge statelode-bridge via npm with STATELODE_TOKEN
- Workspace tokens minted at app.statelode.dev
- Version 0.1.1 from Statelode/statelode repository
- Three MCP tools: task_search, task_get, task_mutate
- Remote MCP https://api.statelode.dev/mcp
- npm bridge package statelode-bridge version 0.1.1
What problem does it solve?
Multiple agent sessions duplicate or contradict todos because there is no shared task API they can search and update.
Who is it for?
Solo builders running several agent workflows on one product who need centralized tasks with MCP-native CRUD-style operations.
Skip if: One-shot chats with no ongoing backlog, or teams unwilling to manage bearer tokens for a hosted task service.
What do I get? / Deliverables
After connecting StateLode with a workspace token, agents read and mutate the same task backlog through three MCP tools.
- Searchable shared task backlog for agents
- Consistent task_get reads across sessions
- task_mutate updates that keep human and agent work aligned
Recommended MCP Servers
Journey fit
Coordinated agent work needs a durable task surface while you build features, not ad-hoc markdown lists that diverge per session. PM subphase covers shared backlogs and task state that humans and agents mutate together.
How it compares
Shared agent task board MCP, not local SQLite memory (Lex) or registry source search (Sourced).
Common Questions / FAQ
Who is StateLode for?
Solo and indie developers using MCP agents who want one workspace task layer instead of copy-pasting todos between sessions.
When should I use StateLode?
Use StateLode during active build and iteration when agents must search existing tasks, fetch details, and update status in sync.
How do I add StateLode to my agent?
Mint a bearer token at app.statelode.dev, add remote MCP https://api.statelode.dev/mcp with Authorization header or run statelode-bridge with STATELODE_TOKEN in stdio config.