Tenjutsu Sébastien Bénard: How the Dead Cells Creator Built a Rogue-Fu Brawler

Why Tenjutsu Sébastien Bénard Is One of the Most Anticipated Indie Action Games

Tenjutsu Sébastien Bénard is a pairing that should get any action-game fan paying attention. Bénard is the lead designer and lead developer behind Dead Cells, a game that sold millions of copies and won Best Action Game at The Game Awards 2018. Now he's building his second solo-dev title, a self-described "rogue-fu" brawler set in an open Japanese city, published by Devolver Digital.

Why does this matter? Because Tenjutsu isn't just another roguelite riding a trend. It's a case study in how a 48-hour game jam prototype evolves into a full commercial release, and how one small design decision can reshape an entire combat system. If you want to understand modern action-game design, the Tenjutsu Sébastien Bénard development story is worth studying closely.

From Ludum Dare to Devolver: The Tenjutsu Timeline

The Tenjutsu Sébastien Bénard journey started with a weekend constraint. The original build was made in 48 hours for Ludum Dare 51, where the theme was "Every 10 seconds." In that version, you fought enemies, collected death marks (skulls), and tagged foes with them. Every ten seconds, marked enemies died. Stronger enemies needed more marks.

That jam entry earned enough positive feedback that Bénard decided to expand it. His own estimate was a six-month project. That was three years ago.

The takeaway for aspiring developers: game jams are not just creative exercises. They're low-risk prototypes that can validate a concept before you commit years to it. Bénard's jam build proved the combat felt good; everything after that was iteration.

The Kunai That Broke the Design Open

Here's where the Tenjutsu Sébastien Bénard story gets genuinely instructive. The early Tenjutsu was a linear beat 'em up built around a finishing move. You'd deplete enemy stamina, but that didn't kill them. Instead, you triggered a special finisher that eliminated every depleted enemy at once. The scoring system rewarded bigger finishers, so you were incentivized to leave weakened, still-dangerous thugs surrounding you.

It was fun, but Bénard identified the flaw himself: the mechanic only worked because of the score and reward loop. It pushed players toward one "perfect run" style, the way stealth games reward flawless ghost approaches. Every other playstyle felt worse. Level design became repetitive because there was essentially one correct way to play.

Then came the fix: a kunai. A throwing blade that could one-shot any enemy, usable once per level. A free joker card for the most annoying target. That single addition opened a floodgate.

• More items felt natural
• Weapons that meaningfully change gameplay followed
• Random weapon drops made each run different
• Random drops led directly to roguelite structure

Bénard had promised himself after Dead Cells that he wouldn't make another roguelite. As he wrote on his dev blog, "Yet here I was, again." The lesson is clear: sometimes the smallest mechanical addition is the one that reveals the game you were actually meant to build.

The Three Pillars Behind Tenjutsu's Combat

Bénard defines game pillars early and keeps them visible, a practice he recommends to any designer. For Tenjutsu, the reference point was the museum combat sequence in John Wick 3: Parabellum. Organic combat built on opportunity, dancing between foes rather than locking onto one, and a wide range of player actions driven by threats in the moment.

That translates into three core design concepts:

Bénard's own example of how this plays: you enter a room with three thugs, one armed with a gun. You chain kicks on the first brawler, elbow the second attacking from behind, finish the first, throw a knife at the gunman, then dash-kick the last thug into a wall. Everyone down in about four seconds. That's the fantasy Tenjutsu Sébastien Bénard is chasing.

Open City, Boss Order Freedom, and the Dead Cells Legacy

One of Bénard's explicit goals was fixing what he saw as Dead Cells' most linear element: world design. In Tenjutsu, you take down bosses in whatever order you choose, and each eliminated boss has consequences for the rest of the run. Later bosses are harder. Don't like the bullet-hell boss? Take him out first. Bénard compares it to Mega Man, but "roguer."

This is a meaningful evolution. Dead Cells' procedural generation used hand-designed room tiles arranged by biome-specific concept graphs, a hybrid approach Bénard documented in detail. Tenjutsu pushes further toward player-directed routing, giving you agency over difficulty pacing rather than handing you a fixed corridor of biomes.

For context on the scale of the legacy he's building from, Dead Cells sold more than 5 million units across platforms. Bénard's sales expectations during early access were around 20,000 units in the first few months; the game moved roughly 100,000 in its first week. That track record is a big part of why Tenjutsu Sébastien Bénard carries so much anticipation.

The Hard Problems Bénard Hasn't Solved Yet

Bénard is refreshingly candid about the open questions in Tenjutsu's development, and they're the same questions any combat designer should be asking:

• Procedural encounter AI: How do you generate dynamic, interesting fights in a procedural setting? Carefully hand-tuning combat in a fixed setup (he cites Sifu as an example he loves) is a very different problem from generating random encounters that stay fresh.
• Readability under pressure: How do you keep one-vs-many combat legible when attacks come from every direction, for players with different attention spans and reflexes?
• Progression: Do players grow through stats, new weapons and techniques, or pure skill? And how do you keep the early game compelling after players have tasted advanced combat?

These aren't marketing talking points. They're the real friction points of blending martial arts choreography with roguelite randomness, and Bénard has said he'll dig into them in future devlogs.

What Developers Can Steal From the Tenjutsu Sébastien Bénard Approach

If you're building your own action game, the Tenjutsu Sébastien Bénard development history offers a repeatable playbook. Here's a practical checklist drawn from his process:

1. Prototype in a jam first. A 48-hour build validated Tenjutsu's core loop before any long-term commitment.
2. Write down your pillars. Two or three concepts, printed and visible. For Tenjutsu: variety, feel, one-vs-many.
3. Watch for the "one correct way" trap. If your scoring system only rewards a single playstyle, your level design will stagnate.
4. Add one wildcard mechanic and see what breaks. The kunai was a small change with systemic consequences.
5. Study game feel obsessively. Bénard released public game-feel demos specifically to show how small details change perceived quality.
6. Reuse your tools. Tenjutsu runs on GameBase, his open-source Haxe/Heaps codebase used across Dead Cells and his jam games.

Player experience with Bénard's earlier work backs up the emphasis on feel. Community reports around Dead Cells consistently praise its responsiveness and the "never the game's fault" death design, which is exactly the standard he's holding Tenjutsu to.

Tenjutsu Sébastien Bénard FAQ

What is Tenjutsu? Tenjutsu is a martial arts roguelite ("rogue-fu") from Sébastien Bénard, the lead designer and developer of Dead Cells. It's set in an open Japanese city where you fight gangs in any order, published by Devolver Digital.

When did Tenjutsu start development? It began as a 48-hour Ludum Dare 51 entry in October 2022. Bénard officially announced the full game in June 2024, and public prototype builds have been released since.

Is Tenjutsu connected to Dead Cells? Not narratively, but Bénard describes it as building on Dead Cells' combat legacy while adding opportunity-based mechanics, crowd control, and non-linear boss order. It also shares his GameBase codebase.

Why did Tenjutsu become a roguelite? Adding the one-use kunai led Bénard to add more weapons, then random weapon drops, which naturally pushed the game toward roguelite structure despite his earlier promise to avoid the genre.

For official announcements and devlogs, follow along at Sébastien Bénard's Deepnight blog, where he posts detailed behind-the-scenes breakdowns of Tenjutsu's development.