Compare Revisions

11 Spirit companions are the bonded battle partners of every playable character in Rewinding Cadence. They are not generic AI followers and they are not interchangeable: each playable is paired with one specific spirit, and that pairing forms the unit that the player controls in the field.
22  
33 Two Companion Archetypes
44  
55 Public material describes two broadly visible visual archetypes, which appear to line up with two different combat styles, though the studio has not confirmed a hard system-level split between them.
66  
77 ArchetypeExamplesCombat FlavorHumanoid spiritsChu, paired with the ReturnerBipedal partners with weapon-based pair attacks; the visuals match a two-character action duo.Animal-form spiritsPaired with several other playables across the castBeast partners with claws, fangs, or elemental attacks; the visuals lean into beast-tamer and ranger silhouettes.
88 Specific named animal spirits beyond the protagonist's pairing have not been published in a stable, central source. As verified material lands the wiki will add per-companion articles with movesets, cultivation tracks, and recruitment routes.
99  
1010 The Pairing is Fixed
1111  
1212 The duo bond is permanent in combat. A playable character cannot trade their spirit companion for a different one mid-cycle, and the spirit cannot be controlled in isolation. This is the foundation that makes duo combat readable: the player learns one pairing's rhythm, gets familiar with the timing of the spirit's interjections, and treats that pairing as the unit of play. Different playables therefore feel meaningfully different, because changing the playable changes the spirit too.
1313  
1414 Spirit Cultivation
1515  
1616 Spirits grow through a cultivation system. The system levels the spirit's base stats and unlocks new techniques over time, and progress in cultivation is one of the things that persists across the time loop, which is what gives long-run progression any traction. Public material has not yet published the full cultivation tree, the exact resource sinks, or the bottleneck materials, so this section will be expanded once verified.
1717  
1818 Confirmed Named Companions
1919  
2020 Only one named spirit companion is publicly confirmed as of mid-2026. The list below tracks the disclosure state of every companion that has surfaced in trailer or beta footage; additional rows will be added when the studio publishes names and kits for the rest of the roster.
2121  
2222 CompanionBonded PlayableArchetypeStatusChuThe ReturnerHumanoid (elf-like)Confirmed; canonical name and pairing published.Unnamed beast companion (urban-city teaser)Unnamed playableAnimal-form (feline silhouette)Visible in trailer footage; official name not published.Unnamed beast companion (mountain-temple teaser)Unnamed playableAnimal-form (avian silhouette)Visible in trailer footage; official name not published.
2323 Per-companion articles will be created only when the studio publishes a stable name and a kit reveal. The wiki will not invent placeholder names or extrapolate movesets from silhouette footage.
2424  
2525 What Cultivation Carries
2626  
2727 Spirit cultivation is one of the few progression vectors that survives the loop reset, alongside the Returner's own growth and certain relationship flags. This makes early cultivation investment the single highest-value habit during a scouting loop, because every gain you make on a companion's cultivation track is still there at the start of the next cycle.
2828  
2929 Carried Across LoopsReset Between LoopsCultivation level on each bonded spiritIn-loop consumables spent on cultivation actionsUnlocked spirit techniquesPer-loop quest flags that gated a specific unlockCultivation milestones earned by clearing fightsDefeated and re-spawning encounter enemies
3030 Specific resource sinks, exact level caps, and per-tier respec rules have not been published in detail. Tester footage suggests cultivation is gated by a mix of materials drawn from open-world activities and milestone unlocks tied to the time loop, but the full economy will be added to this article once the studio publishes it.
3131  
3232 Building a Companion Across Loops
3333  
3434 Because cultivation is one of the persistent layers, a practical first-cycle habit is to schedule a small cultivation investment into every play session even when the loop's main thrust is elsewhere. Small, regular investment compounds, and the alternative is reaching a difficulty wall on a later cycle without the progression to push through it.
3535  
3636 Spend a portion of every loop on cultivation activity. Even a single cultivation block per loop adds up across multiple cycles.Treat companion combat as cultivation practice. Routine fights are the lowest-cost way to surface cultivation milestones for the bonded spirit.Plan cultivation around the calendar. Late-loop days are usually safer for cultivation than early-loop days, when the major story arcs need attention.
3737  
3838 In Story
3939  
4040 Spirits are not just combat units; they are story characters. They speak in field cinematics, react to dialogue choices, and have personal arcs that surface during the dice-gated dialogue branches. Chu in particular drives most of the early-loop story exposition for the protagonist. As more material is published the wiki will add spirit-by-spirit personality notes alongside their combat data.
4141  
4242 Cultivation and the Time Tree
4343  
4444 In the First Playthrough Test (一周目测试) build, the persistent progression layer that anchors spirit cultivation was referred to as the Time Tree (时之树). Unlocking later Time Tree nodes grants one-click level growth to bonded units, and the system does not require duplicate character pulls to advance a unit's effective power. Because the Time Tree itself carries across the loop reset, cultivation gains made on a spirit during one cycle stay live for every subsequent cycle, which is what makes a small but regular cultivation investment in the early loops compound into a noticeable difficulty offset by the later loops.
4545  
4646 The exact Time Tree topology, the per-node cost, and the resource sinks have not been published outside the test build and are therefore not catalogued here. They will be added once the studio publishes them in a public source.