PredictionToolkit logoPredictionToolkit
Language: English
Search tools
← AI agents

Research tool

CloddsBot: configure a local trading assistant and its account permissions

CloddsBot is a self-hosted conversational trading terminal. Its local gateway connects WebChat, messaging channels and AI tools with Polymarket research and configurable execution modules; Kalshi and other configured venues are also in scope.

3.0/ 5Editorial ratingEditorial assessment
Read the CloddsBot repository and setup instructionsgithub.com
Available on
Web · Telegram · Discord · MCP

Screenshots and interface

Product preview1 / 2

The repository’s introduction and WebChat preview.

What is CloddsBot?

For Polymarket work, the documented components include market discovery, portfolio queries, whale tracking and strategy modules. WebChat provides a browser interface, while Telegram and Discord are optional channels into the gateway. Conversation and trading state stay in the configured databases, giving the operator responsibility for the local setup and its logs.

Choose account capabilities deliberately. Research reads and credentialed automation require different configuration, and a messaging integration does not establish identical execution support across venues. The current repository release replaces the unmaintained npm clodds package; runtime compatibility and continuing model or data usage remain part of operating the system.

Who it is for

  • Technical users running a personal trading assistant.
  • Developers connecting conversational interfaces to market tools.
  • Operators needing local control of configuration, state and logs.

Features and coverage

  • Local WebChat

    Use the built-in browser chat with persistent conversations and project organization.

  • Messaging channels

    Connect supported services including Telegram and Discord to the local gateway.

  • Polymarket discovery

    Find recurring crypto rounds and other supported prediction markets.

  • Strategy modules

    Configure supported copying, DCA, arbitrage and other strategy workflows.

  • Risk settings

    Set documented exposure limits, loss limits and kill-switch controls.

  • MCP capabilities

    Expose the bundled tools to compatible AI clients through MCP.

  • Local state

    Store conversation and trading state in the databases you configure.

Workflows with CloddsBot

Start a research assistant

Set up the current repository release with the chosen model provider, then use WebChat to request market information and portfolio reads. This gives the local assistant a research starting point before any separately configured execution modules are introduced.

Add a chat channel

Configure a supported messaging channel during onboarding and verify that its commands reach the intended local gateway.

Operate an automated module

Choose the venue and credentials, set size and loss limits, then inspect the resulting decisions in the trade ledger.

Pricing and total cost

The self-hosted project uses the MIT license. Model-provider API usage, servers, optional paid data services and venue charges are separate operating costs; free source code does not remove them.

Getting started

Before you start: The old npm clodds package is no longer maintained.

  1. Use Node.js 22 or newer.
  2. Install the current GitHub release linked by the repository; the old npm clodds package is unmaintained.
  3. Run onboarding and configure the model provider.
  4. Open local WebChat or connect a supported messaging channel.
  5. Enable the venue modules and account actions needed for your task.
  6. Run the documented diagnostics before leaving the gateway running persistently.

Node.js 20 is unsupported and may fail during dependency installation.

Useful links

    What to check

    • The old npm clodds package is no longer maintained.
    • A supported messaging channel does not mean all venues have the same execution features.
    • Modules with credentials can perform account actions; configure research reads and automated execution separately.
    • Model and data-provider usage can continue for as long as the gateway runs.

    A closer look

    Ratings & reviews

    Editorial rating
    3.0/ 5
    0 user reviews
    PredictionToolkit editorial

    Our take on CloddsBot

    The self-hosted terminal offers flexibility for research, messaging and configured strategies. Runtime requirements, account-capable modules and ongoing provider costs make it a setup the developer or operator must maintain.

    Based on documented features, setup and access terms. How we rate

    User reviews 0

    The main score switches to user ratings after 5 published reviews.

    Be the first to share your experience

    A specific example helps the next person choose the right tool.

    Related tools

    Browse category
    dr-manhattan exchange interface and adapter architecture
    dr-manhattan
    SDK3.5Editorial

    Use shared Python market and order methods across venues, with optional local or hosted MCP access.

    • Language Python
    • Access API · MCP

    A smaller unified Python and MCP integration.

    Simmer SDK introduction and integration paths
    Simmer SDK
    SDK3.5Editorial

    Develop agent strategies with research, paper trading, replay and supported live integrations.

    • Language Python
    • Access CLI

    A strategy development harness with paper and backtest paths.

    PolyClaw OpenClaw market, trading and position commands
    PolyClaw
    Research tool3.0Editorial

    An OpenClaw skill connecting discovery, wallet commands, local positions and logical market-pair analysis.

    • Access CLI

    An agent-oriented Polymarket skill toolkit.

    Frequently asked questions about CloddsBot

    Is it a hosted Telegram bot?

    The core runs on your own machine or server. Telegram is an optional interface to that self-hosted system.

    Can I work in a browser?

    Yes. The gateway includes the local WebChat interface.

    Which installation should I use?

    Use the current release linked by the repository with the documented runtime requirements, rather than the unmaintained npm package.

    What costs remain when the code is free?

    Model calls, hosting infrastructure, optional paid data and venue fees remain separate operating costs.