Tychi Mcp

Created By

yash@tychilabs4 hours ago

Give your AI agent a wallet it can actually use. Tychi MCP plugs into Cursor, Claude Desktop, OpenClaw, and any stdio MCP host — onboard, create, import, and switch wallets with fast direct tools, then balance checks, sends, and policy-gated payments through tyi_chat. Keys live encrypted in ~/.tyi on the operator's machine; a hosted brain parses intent and routes LLM calls while every signature stays local. Includes routing guide (tyi_route), readiness checks (tyi_status), multi-wallet support, spend caps, audit logging, and gasless payment routing via UGF. Built for agent workflows: secrets via tool params or env, never through chat. Apache-2.0, npm + GitHub.

# wallet

# mcp

Overview Content Tools Comments

Overview

Tychi MCP

Give your AI agent a wallet — onboard, hold funds, send, and pay under policy.

Agent-native MCP server for stdio hosts. Fast direct tools for wallet setup and ops; tyi_chat for balance, sends, and policy-gated payments. Keys encrypted in ~/.tyi on the operator's machine. Hosted brain parses intent and routes LLM — signing never leaves the device.

Agents → @tychilabs/tyi-mcp · Humans → @tychilabs/tyi

Architecture

What it does

Agent routing — tyi_route maps intent → correct tool (avoids slow misuse of tyi_chat)
Readiness gate — tyi_status checks mode before any wallet action
Onboarding — tyi_onboard + tyi_onboard_schema for first-time setup (password, LLM provider, API key)
Wallet lifecycle — tyi_create_wallet, tyi_import_wallet, tyi_switch_wallet (fast, no brain loop)
Agent chat — tyi_chat for balances, sends, transfers, payments, policy, limits, history
Multi-wallet — create, import, switch active wallet from one keystore
Policy caps — spend limits enforced before signing
Local signing — private keys in memory on operator machine only
Gasless routing — UGF payment rails for cross-chain gas settlement
Audit log — local activity trail in ~/.tyi
Reset — tyi_reset wipes local data when operator confirms

Agent flow (mandatory)

tyi_route → tyi_status
  → direct tool (onboard / create / import / switch / reset)   ← FAST
  → tyi_chat (balance / send / pay / policy only)              ← SLOW

Never call tyi_chat when tyi_status.ready is false.

Import seed or private key → tyi_import_wallet only (never via tyi_chat).

Tools

Tool	Use when
`tyi_route`	First — intent → tool map
`tyi_status`	Session start — ready? what's missing?
`tyi_onboard_schema`	Field schema before onboard
`tyi_onboard`	First setup or LLM-only setup
`tyi_create_wallet`	New wallet by name
`tyi_import_wallet`	Import mnemonic or privkey (EVM / Solana / Sui)
`tyi_switch_wallet`	Change active wallet
`tyi_reset`	Wipe `~/.tyi` (`confirm: true`)
`tyi_chat`	Balance, send, pay, transfer, policy, limits

Install

npx -y @tychilabs/tyi-mcp@beta
npx @tychilabs/tyi-mcp@beta --tools

MCP host config

{
  "mcpServers": {
    "tychi": {
      "command": "npx",
      "args": ["-y", "@tychilabs/tyi-mcp@beta"],
      "env": {
        "TYI_PASSWORD": "<from tyi_onboard>",
        "TYCHI_BRAIN_URL": "http://hosted_brain.tychilabs.com"
      }
    }
  }
}

Onboarding modes

Mode	Meaning
`fresh`	No wallet — run `tyi_onboard`
`llm_only`	Wallet exists, no LLM key — onboard LLM fields only
`ready`	OK for `tyi_chat`

Operator provides (never invent): keystore password, LLM provider + API key, optional mnemonic for import.

Set TYI_PASSWORD in MCP host env after onboard — not in chat.

Environment

Variable	Required	Default
`TYI_PASSWORD`	After onboard	—
`TYCHI_BRAIN_URL`	No	`http://hosted_brain.tychilabs.com`
`TYI_DATA_DIR`	No	`~/.tyi`
`KEYSTORE_PASSWORD`	Alias	same as `TYI_PASSWORD`

License

Apache-2.0 — see LICENSE

Try in Playground

Server Config

{
  "mcpServers": {
    "tychi": {
      "command": "npx",
      "args": [
        "-y",
        "@tychilabs/tyi-mcp@beta"
      ],
      "env": {
        "TYI_PASSWORD": "<keystore-password-after-tyi_onboard>",
        "TYCHI_BRAIN_URL": "http://hosted_brain.tychilabs.com"
      }
    }
  }
}

Project Info

Created At

4 hours ago

Updated At

4 hours ago

Author Name

yash@tychilabs

Star

Language

License

Recommend Servers

View All

Genflik

@GenFlik

Create professional AI talking-avatar marketing videos from a script or product idea — pick an avatar and voice, add product screenshots or AI B-roll, and render a downloadable ad. Built for e-commerce , Amazon & Shopify sellers.

21 hours ago

Github

@modelcontextprotocol

Repository management, file operations, and GitHub API integration

a year ago

Video Overlay Kit

@alichherawalla

AI-driven animated b-roll overlay renderer for short-form video. Paste your script into your AI coding tool, the MCP server writes the scene spec and renders an mp4. Free, MIT, local.

a day ago

Dochost

@zyli5313

Publish Markdown or HTML to a clean, shareable link straight from your AI assistant. Hosted remote MCP server (Streamable HTTP) with OAuth sign-in — no API keys. Two tools: publish (Markdown/HTML → URL, with optional password, custom slug, and unlisted/public) and list_my_pages. Connect: claude mcp add --transport http dochost https://dochost.io/api/mcp . Homepage: https://dochost.io

13 hours ago

4 hours ago

Toksta MCP lets you run creator discovery and analysis workflows inside AI clients like Claude and ChatGPT, using your Toksta account, campaigns and credits. Search LinkedIn and YouTube creators, inspect real post evidence, enrich with engagement metrics, run content-fit and audience-fit analysis, and save shortlists back to Toksta lists.

7 hours ago

Snapforge

@sporty303

Screenshot & HTML-to-PDF rendering API for AI agents — capture any URL or raw HTML as PNG/JPEG/PDF via a managed Chromium fleet. Free tier.

21 hours ago

Numox

a day ago

302_browser_use_mcp

@302ai

Automatically create a remote browser to complete your specified tasks, developed based on Browser Use + Sandbox. 自动创建一个远程浏览器，完成你指定的任务，基于Browser Use + Sandbox开发。

a year ago

MCP Advisor

@istarwyh

MCP Advisor & Installation - Use the right MCP server for your needs

TypeScript

a year ago

Geniuz

@jackccrawford

Open-source MCP server that gives Claude persistent memory. Your AI keeps your standards, decisions, and reasons across sessions. Recalled by meaning, not keywords. Runs fully local: on-device semantic search, no cloud, no API keys. macOS, Windows, Linux. MIT.

16 hours ago

Dependency Freshness Checker

@Armigerous

**Dependency Freshness Checker** tells any AI coding agent whether an npm or PyPI package is **outdated (out of date)** — and gives the cited facts to prove it: the latest version, release dates, deprecation status, how many versions behind you are, and a dated **"what changed since your version"** breaking-change diff. It is MCP-native, reads only public registries and GitHub releases (no scraping, no ToS risk), and is priced Pay-Per-Event for pay-as-you-go agent use.

2 days ago

Zoviz Mcp

8 hours ago

Mcp Server Chatsum

@chatmcp

summarize chat message

typescript

a year ago

Rierino Documentation

7 hours ago

Rami Code Review

@Shavakan

Rami reviews every pull request for the bugs, logic errors, and risky patterns that AI coding agents commonly leave behind. When it flags something, your agent — Claude Code, Cursor, or Codex — fetches the finding over MCP, fixes it, and pushes; Rami re-reviews until those issues are cleared. A passing Rami check means the PR is free of those common agent flaws — not that it's ready to merge; that call stays yours. Disagree with a finding? Dispute it with evidence and Rami re-evaluates. Free to start — 2 reviews/day, no credit card.

a day ago

Trainzilla Mcp

2 days ago

Amazingmcp — Mcp Server For .net / C# Codebases

@remleo

An MCP server that gives AI agents deep understanding of C# codebases via Roslyn — type search, dependency graphs, usage analysis, and architecture overviews, all from a live in-memory compilation.

25 minutes ago

a day ago

12 hours ago

Generate, edit, and upscale AI video & images from any agent — VicSee as a set of MCP tools. Multi-model (Seedance, Veo, Kling, FLUX, Nano Banana), text or image to video or image, with async task polling. Six tools: list models, generate, poll task, upscale image, upscale video, check credits.

30 minutes ago

Inferbench

@JoniMartin27

InferBench's MCP server lets coding agents run, serve and benchmark local LLMs (text + image, llama.cpp + Stable Diffusion) on your own hardware on demand. Measures real tokens/sec, picks the optimal quant for your GPU, and exposes a 124-model catalog. Local-first, no cloud required.

3 hours ago

Lorekeeper Mcp

@Jessinra

2 hours ago

Clavio Global Data Resolver

@y-yahiaoui1406

Real-time data resolver for autonomous agents. 1M+ endpoints covering 200 countries, 169k cities, economic indicators, FX, weather, biodiversity, seismic data. Africa/MENA depth. x402 payments on Base Mainnet. Status : created

19 hours ago

Blame Today

@royashbrook

Blame things on blame.today, the anonymous "who do you blame today" board. Nostr, no account.

14 hours ago

Tatsu55

a day ago

Antics

@antics-gg

An MCP server that lets an AI agent deploy a web game to a playable multiplayer URL — rooms, live state sync, and leaderboards — in one conversation.

9 hours ago

Mcp Office Powerpoint

@dosev-ai

Local-first MCP server for Microsoft PowerPoint. 46 tools include slide CRUD, shape and text editing, an Output Contract framework for machine-verifiable slide specs, evidence bundles, table editing, and image insertion. File-based tools (python-pptx) run on Windows without PowerPoint installed. COM export tools (PDF, PNG, slide show) additionally require Microsoft PowerPoint. No hosted tier — runs on your machine.

a day ago

Puppeteer

@modelcontextprotocol

Browser automation and web scraping

a year ago

Zambo Ai Tools

10+ advanced tools for AI agents

a day ago

Tychi Mcp

Tychi MCP

What it does

Agent flow (mandatory)

Tools

Install

MCP host config

Onboarding modes

Environment

Links

License

Server Config

Recommend Servers