Your money stays yours
We have no wallet, no pooled balance, no deposit address. Funds sit in your exchange account and orders are placed there under your name.
About Candleweb
Candleweb is infrastructure for agentic trading, built in Africa. We started because the interesting half of an automated trading system - the part that decides - had become easy, and the boring half had not. Someone had to do the boring half. We decided it should be us, built here rather than imported.
How we got here
ThenA trading bot, aimed at people who wanted one.
↓NowInfrastructure, aimed at people who build their own.
Candleweb began as a trading model. We built a reinforcement learning agent, trained it, connected it to an exchange, and watched it work - right up until a real order half-filled and the whole thing fell over.
That failure taught us more than the model did. The intelligence was never the constraint. The constraint was everything downstream of the decision: watching a position at three in the morning, enforcing a stop when the connection drops, and reconciling a fill that arrived in four pieces.
Then we taught a room full of people to build agents. They built things that formed genuinely good opinions about the market and got stuck at exactly the same wall we had. That was the moment Candleweb became infrastructure. You bring the strategy. We run everything after it.
What we hold to
We have no wallet, no pooled balance, no deposit address. Funds sit in your exchange account and orders are placed there under your name.
We store the exact candles and indicator values your bot saw at the moment it chose, next to what it decided and what happened after.
We built for the people in our own classrooms first, and what came out of that works for a builder anywhere.
Two ways in
Either way an agent does the trading. Candleweb runs the same infrastructure underneath both routes - what changes is who writes the strategy.
We feed your n8n, Make, or code agent live market data, take back its decision, and run the trade from there.
See how it works ↗Our own trading agent reads the market, makes the call, and runs the position on the same execution and monitoring layer.
Explore CandleAI ↗Same execution path. Same monitoring. Same trade records. Same rule about who holds your money - you do.
Who we work with
Anyone can call a public exchange endpoint. We went further, because people trusting us with their trading deserve to know we have supported relationships with the venues their orders reach.
A signed broker partnership and supported, accountable channel for agent orders.
Your agent behaves identically whichever supported venue you trade on.
Licensed local currency rails across Africa for funding and withdrawals.
The road so far
Founded to bring automated trading to African retail investors.
Partial fills, rate limits and position monitoring turn out to be the real work.
Students build working agents and hit the same wall we did.
Tutor's automation track is recognised by n8n.
Candleweb repositions from a trading product to infrastructure.
Market data, monitoring, execution and review between a decision and a closed position.
The people
We're deliberately small, and most people here sit across more than one of these. What follows is how the work is divided.
Market structure, indicators and regimes. Keeps CandleAI's strategy honest.
Agent architecture, model work and the questions we have not answered yet.
The platform, API, workers and everything that must still be correct at 3am.
Tutor's cohorts and curriculum take someone from no automation experience to a working agent.
How Candleweb is explained, and the review every claim about money passes.
Tutor by Candleweb
Tutor is our cohort programme in AI automation and personal finance. Students learn APIs and webhooks, build real workflows in n8n, and think properly about money before trading it.
It exists because the people best served by Candleweb are people who can build. Rather than wait for them to appear, we teach them.
Explore courses ↗Work with us
We hire rarely and slowly, and care more about what you've built than where you studied. If infrastructure that has to be correct at three in the morning appeals to you, we'd like to hear from you.