Steam Next Fest 2026 Demo Short Description - 300 Character Evidence Checklist for Micro-Studios

Your short description is not a devlog hook. It is a high-velocity contract Steam repeats beside your capsule in browse surfaces, tag intersections, and recommendation rows. During October 2026 Next Fest traffic, players skim that contract in under two seconds. If a single clause promises co-op, full-game scope, or a platform your fest demo does not prove in hour one, you are not “bad at marketing”—you are one email away from the same store-demo mismatch yellow pattern we documented in the 48-hour recovery case study.

This tutorial is for micro-teams who already respect builds but still treat copy as soft art. You will get a 300-character budget you can defend in a diff, a checklist aligned with Wednesday metadata ritual exports, and rewrite patterns that survive mock audit dimension seven without inventing metrics or fake studios.

Time to read: ~28 minutes. Time to execute first pass: ~90 minutes with exports already wired.

Why this matters now (May–October 2026)

1. Steamworks still caps short description length — the limit is not a suggestion; exceeding it fails paste workflows and trains teams to “shrink in Steamworks” without a git baseline (see official written description documentation).
2. Next Fest compresses attention — discovery rows show short description + capsule together; contradictory clauses read louder than long-form About sections.
3. Partners diff faster than players — intake compression means reviewers paste your short description next to build_id screenshots the same afternoon.
4. LLM drafts default to hype density — short fields tempt “make it punchy” prompts; you need the same FAQ LLM human diff gates discipline, compressed into 300 characters.
5. Metadata tooling is mainstream — if you still cannot diff, use fourteen free export tools; lack of tooling is no longer an excuse in Q2 2026 planning rooms.

Direct answer: Treat the short description as G1 identity copy in your weekly export, count characters against Steam’s cap, cold-install the demo, delete any clause not experienced in hour one, then attach the diff row to your release-evidence log before you promote a branch.

Who this is for and what you ship

Prerequisites (tools and habits)

Short description vs About section vs capsule story

Short description errors are worse during Next Fest because they ride along capsule pixels you tuned with 32px icon gates and screenshot safe zones. Visual truth plus textual lie equals refund tone, not curiosity.

The 300-character budget (line anatomy)

Think in slots, not prose vibes.

Demo boundary language is non-negotiable for Next Fest builds. If you refuse the word “demo,” you must still signal slice scope (“first chapter,” “vertical slice,” “combat sandbox”) that matches hour one.

Hard rails (do not negotiate)

• No platform list unless demo binaries exist there now.
• No multiplayer verbs unless session code or split-screen exists in the fest zip.
• No roadmap verbs (“soon,” “coming”) — those belong in announcements, not short description.
• No price or discount language — keep commerce in regional pricing worksheets and long-form sections.

Evidence checklist (twelve gates before paste)

If any gate fails, you are not “almost ready.” You are one browse impression away from a player who screenshots your sentence beside a crash or missing mode.

Ten rewrite lenses (micro-edits that save macro trust)

1. Verb downgrade — change “master” to “learn” unless mastery is literally the tutorial beat.
2. Scope shrink — replace world-spanning nouns with the single biome in the demo zip.
3. Co-op excision — delete “with friends” unless invite flow ships.
4. Platform list cut — keep only the binary you upload for fest.
5. AI clause honesty — if procedural dialogue uses models, say so in long-form; keep short description on gameplay loop.
6. Remove comparative adjectives — “best,” “only,” “first” unless legal reviewed.
7. Swap lore for interaction — one proper noun max unless title itself is unknown.
8. Replace roadmap with boundary — “Act 1 slice” beats “full saga preview.”
9. Tone match capsule — if capsule is grim, do not promise cozy farming.
10. Ask the vertical slice question — seven-day vertical slice challenge Day 4 store parity applies to this field verbatim.

Each lens should take under five minutes once your team names a rotating “lens owner” during the five-day metadata freeze sprint.

Wednesday ritual integration (where short description lives)

Add short description as an explicit export line in your Wednesday Steam metadata diff packet, not as an afterthought when someone notices browse text. Field groups G1–G3 already cover identity surfaces; short description is effectively G1 compressed. If your exporter cannot capture it yet, paste manual short-description.txt until tooling catches up—empty diff beats heroic memory.

Sample ninety-minute short-description sprint

Truth audit alignment (week-scale, not fest eve)

Run the seven-day wishlist truth audit when your short description has changed more than twice in a month. Short description is day-one material in that audit—do not skip it because About section feels “more important.” Search surfaces disagree.

Partner and diligence readers (tone they expect)

Same text, different paranoia—one lie fails all three at once.

LLM drafting discipline (short field, full gates)

If you use LLMs, never paste raw output into Steamworks. Pipeline:

1. Prompt with hard character cap + bullet list of banned clauses.
2. Generate three variants.
3. Engineer marks each sentence pass/fail against hour one.
4. Writer merges only passing sentences.
5. Run Gate D from FAQ LLM pipeline adapted for length (same honesty, shorter field).

Next Fest cadence table (May through October)

Pair with Asia–EU handoff evidence if your writer and builder clocks do not overlap—short descriptions love to drift at 2am exports.

Multiplayer and platform traps (expanded)

Multiplayer

If netcode exists internally but not in fest zip, the short description must read single-player or local only honestly. “Online later” belongs in news hub posts, not this field.

Platforms

If you only ship Windows demo, delete macOS or SteamOS verbs. Deck compatibility claims belong next to actual Deck testing evidence, not beside wishful thinking.

Early Access and live-service language

Early Access games still ship a slice. Short description must describe the slice players install this week, not the 2027 finale. If roadmap optimism sneaks in, move it to long-form roadmap sections you actually maintain.

Commerce cross-check (G5 bleed)

Short description is not a price field, but writers sometimes drag in DLC or deluxe nouns. If you mention cosmetics, sync with DLC anchor worksheet and G5 commercial rows in Wednesday exports so you do not create a second lie class.

Crash spikes and copy innocence

If your demo crashes on common GPUs, fix crash-log discipline before you claim stability adjectives. Short description is not a crash report, but “stable” and “polished” are empirical claims—downgrade verbs until evidence returns.

UTM and attribution hygiene

When you run twenty-one-day UTM experiments, tag campaigns with honest landing expectations. Misleading short descriptions poison attribution—you will mis-read which channel brought angry wishlists.

Two-storefront rule

If you follow the two-storefront fragmentation rule, diff short descriptions per surface. A Steam sentence that matches demo but contradicts itch first paragraph still erodes trust for cross-linkers.

Evidence cycles vs weekly hype edits

Evidence cycles opinion applies: batch non-critical copy edits weekly. Short description should not change daily unless you are fixing a provable lie—oscillation trains algorithms and humans that your words are noise.

Friday Block 5 maintenance hook

Add a single checkbox to Friday Block 5: “Short description diff green?” Fifteen seconds if exports exist; hours if you skipped Wednesday all month.

Operating review tie-in

If you run four Friday five-block operating reviews, put short-description incidents in Block 2 under “store truth.” Producers love moving that block fast when the field is quantified to 300 characters.

Hash and manifest sanity (when partners distrust bytes)

If your yellow is actually artifact confusion, not English, pivot tools from SHA256 manifest drill and hash tool listicle—but still fix short description if it references modes missing from the zip.

Worked example (anonymous fantasy roguelite slice)

Before (fails hour one): “Drop into a living city with friends, master four schools of magic, and decide the fate of three factions in this narrative roguelite—full voice cast, Steam Deck verified, mod-ready.” Why it fails: co-op not in demo, only one magic school appears in tutorial beat, factions are not choosable in slice, voice cast is partial, Deck verification still pending, mods not exposed.

After (passes hour one): “Next Fest demo — third-person roguelite slice: learn your first school’s combos, die, unlock meta upgrades, repeat. Single-player; Windows build; full VO in tutorial only.” Why it passes: verbs match tutorial loop, boundary explicit, platform singular, VO scope honest.

This rewrite pattern is boring on purpose—boring survives screenshots in partner ZIP diligence threads.

Cold install pairing (engineer script notes)

Ask engineering to paste a five-line sign-off beside your git PR:

1. Steam account used (internal tester).
2. Branch name and build_id.
3. First quest completed or first fail state reached.
4. Any skipped tutorial steps (must be “none” if short description claims full tutorial).
5. GPU and driver note if you claim performance comfort.

That stub costs ten minutes once per promotion and prevents the “writer fixed copy while engineer promoted yesterday” drift pattern that triggered store-demo mismatch recovery cases.

When marketing fights for “bigger” language

Marketing will ask for emotional height—give them capsule art, trailer middle beats, and long-form About poetry. Negotiate that the short description stays a spec sheet in prose clothing. If they need a rally sentence, put it in social posts where you can edit hourly without implying binary parity.

Snippet-friendly answers

What is the Steam short description limit?
Steam caps the short description field for store pages; treat 300 characters as the planning budget your team drafts against before paste—verify against current Steamworks guidance linked above.

Does Next Fest require a unique short description?
No special field—but discovery density makes honesty cheaper than refunds.

Is short description the same as the first About paragraph?
No—different surfaces and editing rhythms; diff both weekly.

Can I mention co-op if it ships after fest?
Not in this field for the fest demo slice; use announcements.

Do emojis count toward the limit?
Assume yes—count everything you paste.

Should I localize short description day one?
Only if localized demos exist; otherwise mark single-language honestly per localization tooling article.

What if I am over by three characters?
Cut an adjective, not the demo boundary clause.

Who approves final paste?
Producer plus engineer co-sign high-risk verbs; writer owns tone.

How often should this change during October?
Ideally never except typo fixes; branch changes force re-run gates.

Does this replace long-form About work?
No—this field is the compressed trailer sentence for browse contexts.

Common mistakes (ten)

1. Treating Steamworks as canonical text storage.
2. Letting marketing tweak without cold install.
3. Listing engines (“Made in Unreal”) instead of player verbs.
4. Hiding “demo” language to look “bigger.”
5. Copying competitor sentence rhythm with incompatible scope.
6. Forgetting Deck glyph claims when not tested.
7. Mismatched tone with capsule palette lock.
8. Using LLM output without engineer pass.
9. Changing short description during ad spend spikes without freeze.
10. Skipping diff because “it is only 300 characters.”

Pro tips (eight)

1. Keep short-description.md beside faq-approved/ in git.
2. Add wc -m or PowerShell Measure-Object -Character to CI optional lint.
3. Tag releases store/short-desc-2026-10-01 when locking fest copy.
4. Pair edits with prompt registry freeze if runtime AI also changed.
5. Screenshot browse surfaces monthly—visual context matters.
6. Teach interns the twelve gates before Steamworks access.
7. Mirror short description into partner README bullet two for zip packets.
8. Log changes in release-evidence taxonomy under 03-store/.

Key takeaways

• Short description is a browse-surface contract, not a mini devlog.
• Plan against the 300-character cap before Steamworks paste; verify with official Steamworks docs.
• Align wording with fest demo hour one, not roadmap fantasy.
• Add the field to Wednesday metadata exports as G1-adjacent truth.
• Run twelve evidence gates before treating copy as shippable.
• Use ten rewrite lenses for fast de-risking without bloating length.
• LLM drafts still need engineer pass + Gate D honesty.
• Multiplayer and platform lists are the highest-risk clauses—default narrow.
• Next Fest cadence should freeze this field by September unless build_id forces updates.
• Pair with truth audit, vertical slice Day 4, and store-demo mismatch recovery patterns for end-to-end parity.
• Commerce and DLC language belongs in G5 long-form checks, not squeezed into 300 chars.
• Treat Friday Block 5 checkbox as cheap insurance against drift.

FAQ

Can short description mention EA roadmap?
Only if sentences describe what exists today; move future tense to news posts.

What if Steamworks UI counts differently than VS Code?
Paste into Steamworks preview first; export immediately after save.

Should students solo devs skip git?
No—use a single markdown file in repo even if solo.

Does Steam strip formatting?
Assume plain text; do not rely on markdown rendering in short description field.

What about adult content gating language?
Follow Steam guidelines in long-form; short description should not promise unrated modes.

Can I mention mod support?
Only if demo ships mod tools or documented sample mods in zip.

What if our game is narrative-heavy?
Name one interaction verb plus emotional tone without spoiling late acts.

How do we train contractors?
Give them this checklist PDF; remove Steamworks paste rights until gate twelve passes once supervised.

Is there an SEO trick density?
No—honest genre words outperform stacked buzzwords for trust and refunds.

What if we ship hotfix mid fest?
Re-run cold install + update short description only if hour-one behavior changed.

Can we A/B test wording weekly?
Only with evidence discipline—otherwise you violate evidence-cycle guidance and confuse returning players.

Where do press quotes go?
Long-form About or trailer end cards—not this field.

Conclusion

Three hundred characters is small enough to edit in a waiting room and large enough to destroy trust when it disagrees with the bytes players download. Treat the short description as part of your ship evidence, not your mood board. Lock it with the same seriousness you lock build_id, run it through Wednesday diffs, and let October traffic hit a store page where browse text, capsule art, and demo scope all tell one honest story.

If you only remember one sentence after Next Fest prep: the short description must survive being screenshot beside your depot id—because in 2026, it will be.