Agent Builders Cup
Resources and Rules
The official rules, plus everything you need to build your agent — videos, guides, examples, and support.
Official Rules
The rules of record for this competition. Expand any article for the full text.
What You Can Build
Videos, guides, code, and a real example for each of the two eligible paths.
Build a trading strategy that trades on one of the sponsoring venues. You can build it either as a Hummingbot V2 Controller — rule-based logic that runs natively inside Hummingbot — or as a Condor Agent, an LLM-driven agent that reasons about live market conditions. Both race on equal footing, so pick the abstraction that fits the strategy you want to run.
Building Controllers
Rule-based decision logic — when to enter, exit, or adjust.
Understanding Market Making Controllers (Part 1)
Federico breaks down how V2 market making controllers work, live.
Understanding Market Making Controllers (Part 2)
Continuation — deeper into controller configuration and tuning.
V2 Controllers — Overview
What a v2 controller is and how it fits into Hummingbot.
V2 Controllers — Walkthrough
Step-by-step guide to building your own controller.
How to Configure a V2 Strategy Controller
Setting up a controller in Hummingbot, step by step.
Strategy Guide: Grid Strike
A grid-based controller with dynamic risk management.
Controller Strategy Examples
Real submissions to study — full code and strategy.md.
@raj/cohort-13-market-making-at-the-touch
Market Making At-The-Touch
A market-making strategy optimised for posting at the top of the order book (the "touch") on both sides. It decides, at every tick, whether to keep a buy LO at the best bid and/or a sell LO at the best ask, based on the agent's current inventory and how much time is left in a rolling horizon. The decision rule comes from solving an offline stochastic-control problem; the live controller is just a fast lookup against a pre-baked table.
@vzmievski/cohort-13-pmm-mister-dynamic
PMM Mister Dynamic
PMMMisterDynamic is a modification of PMM Mister strategy where some of original PMM Mister static parameters are replaced with a dynamically adjusted functions of the market volatility and trend
@isreallee82/cohort-13-flowedge
FlowEdge
Regime-adaptive directional controller with multi-timeframe signal generation, NATR-scaled DCA entries, funding rate bias, and embedded self-adaptive threshold. Uses ADX regime detection to suppress entries in choppy markets and fire 3-level maker DCA positions when both fast (3m) and slow (15m) timeframes confirm a trending regime.
Building Agents
An LLM-driven agent that reasons about market conditions and decides what to do.
Introducing Condor
The open-source harness connecting LLMs to Hummingbot's trading infrastructure.
Creating AI Trading Agents in Condor
Live walkthrough of building an autonomous trading agent in Condor.
The Bot Pod Ep. 1 — Trading Agents in Condor
Hummingbot's weekly livestream — building an agent from scratch.
The Bot Pod Ep. 11 — Agent Builders Cup Kickoff
Hackathon kickoff session, F1-style finals, and Condor sub-agent architecture.
Condor Docs
Framework docs — setup, architecture, and building your first agent.
Agent Strategy Examples
Real submissions to study — full code and strategy.md.
@james-lo-tsz/cohort-13-meteora-lp-agent
Meteora LP Agent
Autonomous concentrated liquidity market-maker on Meteora Dynamic Liquidity Market Maker (DLMM) on Solana. Each tick, it scans Meteora DLMM pools via GeckoTerminal, selects the pool with the highest volume-to-total-value-locked (V/T) ratio that passes structural and routability checks, rebalances the wallet via Jupiter if needed, and deploys a managed liquidity-providing (LP) position — closing it automatically when defined exit conditions are met.
@hummingbot/agents/market_making_expert
Market Making Expert
Market making specialist — regime detection, spread calibration, and inventory management for PMM strategies. Ships with Condor as a reference agent.
Exchange Sponsors
The venues fielding teams in this competition.
To maximize your chances of being selected, your strategy should ideally trade on a sponsoring exchange. Each team picks the agents and controllers that best fit its own venue.
Eligible Exchanges
Agents may trade on any of these venues during the race.
Connector Guides
How to configure each eligible venue's connector in Hummingbot.
Binance
Connector setup, API keys, and supported order types.
Gate.io
Spot & perpetual connector setup and API keys.
Hyperliquid
Wallet & API key auth, spot and perpetual connectors.
Derive
Spot & perpetual connector setup for the onchain venue.
XRPL
Connector setup for the XRP Ledger's onchain order book.
Solana
Gateway setup for Solana DEXs — Jupiter, Meteora, Orca, and Raydium.
Bitget
Spot & perpetual connector setup and API keys.
Hackathon Timeline
The key dates of the competition, from registration through the finals.
Registration
May 1 – Aug 15, 2026NowSign up, follow the sponsor workshops, and apply to the teams you want to race for.
Hackathon
Aug 1 – Aug 31, 2026Build your trading agent across the build window and submit before submissions close.
Judging
Sep 1 – Sep 30, 2026Botcamp validates strategy code and sponsors submit final rankings to pick their agent drivers.
Finals
Oct 1 – Oct 2, 202648-hour livestreamed competition. Winners announced Oct 7 at our Token2049 side event in Singapore.
Winners are announced on October 7, 2026.
Getting Support
Reach the Botcamp team and other builders during the race.
Join the #hackathon channel on Discord
Talk with other builders, ask the Botcamp team questions, and get help debugging your agent — all in the Hummingbot Discord.