name

neta-travel

description

After adopting a character (or on heartbeat/manual trigger), take the character on a Neta world adventure — find a matching world, explore curated scenes, generate one personalized image per stop, collect 5 stops to unlock an interactive travel atlas. **Auto-trigger:** - After adopt completes and user clicks "Explore this world" - User mentions: travel / explore / let's go / wander / random - User says "more" / "continue" in a travel image context **Prerequisites:** SOUL.md with character name + portrait image. No external dependencies. **Stream feedback after each step — don't batch.** **Speed:** Push forward immediately. If image poll >30s, output a waiting message. **Language:** Match trigger language. Stick to it unless user switches. **Interaction:** All next-step choices as clickable quick-reply buttons. Click sends `@{bot_name} {option}`. **Output:** Character dialogue and images output directly. Narration/system messages use discord plugin.

Neta Travel

Take a Neta character into a world that matches their soul — explore curated scenes, generate personalized travel images, collect 5 stops to unlock an interactive atlas.

---

Setup (silent, first run only)

travel.js: Check ~/.openclaw/workspace/skills/neta-travel/travel.js. If missing:

mkdir -p ~/.openclaw/workspace/skills && cd ~/.openclaw/workspace/skills
git clone https://github.com/talesofai/neta-travel.git && cd neta-travel

Token: Check NETA_TOKEN env → ~/.openclaw/workspace/.env → ~/developer/clawhouse/.env. If missing:

⚠️ NETA_TOKEN required. Add it to ~/.openclaw/workspace/.env

All commands below run as node ~/.openclaw/workspace/skills/neta-travel/travel.js <cmd>. Shorthand in this doc: travel <cmd>.

Adopt (if no SOUL.md or no portrait image):

travel adopt "<character_name>"

Looks up the character in the TCP database, generates a portrait image, and writes ~/developer/clawhouse/SOUL.md. Custom path: travel adopt "<name>" /path/to/SOUL.md

---

Language

Detect from trigger word: Chinese → Chinese, English → English, other → follow user. Switch only if user explicitly requests it.

---

Steps

Output feedback after each step immediately.

Step 0 · Adopt (if no SOUL.md, silent)

travel adopt "{character_name}"  →  { name, picture_uuid, soul_path, image_url }

If SOUL.md is missing or has no 形象图片, run adopt first. Searches TCP for character, picks an existing portrait from the character's image feed (instant — no generation), writes SOUL.md automatically. ~2s.

Step 1 · Read character (silent, <1ms)

travel soul  →  { name, picture_uuid }

Reads SOUL.md. Extracts 名字 → character_name, 形象图片 UUID → picture_uuid.

⚠️ Missing 形象图片 = no reference image = generation FAILURE. Run Step 0 first.

Step 2 · Generate world context (LLM, no API)

Based on character's personality, background, and tags from SOUL.md, infer:

• world_name — a Neta world that fits the character's soul
• world_description — 2–4 paragraphs: setting, rules, atmosphere, character fit
• world_count — use 500+ (no API needed)

Step 3 · Opening (segmented output)

Segment 1:

╔══════════════════════════════════════════════╗
║           N E T A   U N I V E R S E          ║
╚══════════════════════════════════════════════╝

  Worlds mapped: {world_count}
  Every world holds a story waiting to happen.

Segment 2:

  ────────────────────────────────────────────
  Scanning soul frequency for {character_name}...
  ████████████████████████  Match found
  ────────────────────────────────────────────

Segment 3: Output world_name reveal, then world_description paragraphs one by one:

  ╔══════════════════════════════════════════╗
  ║  World: {world_name}                     ║
  ╚══════════════════════════════════════════╝
  {world_description_paragraph_1}
  {world_description_paragraph_2}  ...

Segment 4:

  ════════════════════════════════════════════
  {character_name} and this world share an inexplicable gravity.
  ════════════════════════════════════════════

Button: Explore 🌀 → @{bot_name} explore

---

Exploration (after user clicks Explore)

Maintain visited_uuids list in memory throughout the session.

Step 4 · Find destination (~1s)

travel suggest {visited_uuids joined by comma}  →  { uuid, name }

Calls recsys recommend API, excludes visited UUIDs. Falls back to interactive feed if empty.

Output:

🌀 Portal opening...
📍 Destination locked: {name}

Step 5–8 · Generate image (30–60s)

travel gen "{character_name}" "{picture_uuid}" "{uuid}"
→  { scene, status, url, collection_uuid }

Internally:

1. Fetches scene prompt template from collection
2. Resolves character TCP UUID → builds oc_vtoken_adaptor vtoken for precise character binding
3. Submits to 8_image_edit model with character reference image
4. Polls until SUCCESS / FAILURE

Output as each stage completes:

🔍 Scene loaded: {scene}
🎨 Painting the scene...
⏳ Rendering is taking a bit longer, almost there...  ← only if >30s

On FAILURE: ⚠️ Lost the way this stop — try a different destination?

---

Station Display

On SUCCESS, output in this order:

1. Character scene (before image):

🎭 [{scene}]

{1–2 sentences: environment and atmosphere on arrival}

{character_name}: {first-person reaction matching SOUL.md personality and speech style}
({action/expression, 1 sentence})

2. Header + image (URL on its own line — auto-embeds):

━━━━━━━━━━━━━━━━━━━━━━━━
Stop {round} · {scene}
{url}

3. Progress + add collection_uuid to visited_uuids:

Stop	Bar	Message
1	▓░░░░ 1/5	🌟 Stop 1! 4 more to unlock your atlas. Keep going?
2	▓▓░░░ 2/5	✨ Two in! Atlas is getting closer~
3	▓▓▓░░ 3/5	🔥 Halfway! Two more and the atlas is yours!
4	▓▓▓▓░ 4/5	⚡ One stop away! Atlas within reach!
5+	▓▓▓▓▓ 5/5 🎉	Atlas unlocked! Type "generate atlas" or keep exploring~

4. Buttons:

< 5 stops: Continue 🗺️ → @{bot_name} continue · Call it a day 👋 → @{bot_name} end

≥ 5 stops: add Generate Atlas 📖 → @{bot_name} generate atlas

---

Travel Atlas (user-triggered)

Triggers on: "generate atlas" / "atlas" / "gallery" / "html"

Ask for style preference:

What style for the atlas? (skip = default map)
e.g. retro film / starmap / pixel game / minimal white...

Default — interactive map: Each stop's image as a landmark on an adventure-style map (parchment / pixel / star chart). Click to expand image + scene name + stop number. Match map style to character vibe.

Custom: Any layout (gallery, card wall, timeline, magazine). Keep click-to-expand.

Save to ~/.openclaw/workspace/pages/travel_{character_name}_{date}.html. Ask for username once per session → output share link on its own line:

🔗 https://claw-{username}-pages.talesofai.com/travel_{character_name}_{date}.html

Offer re-customization: Want a different style? Just describe it ✨

---

Errors

Error	Cause	Fix
SOUL.md not found	adopt not run	travel adopt "<name>"
task_status: FAILURE	Missing/invalid picture_uuid	Re-run adopt to get valid portrait
Character not found in TCP	Wrong name or not in Neta DB	Try exact CN/EN name
No existing portrait found	Character has no public images	Try a different character name
code 433	Concurrent limit	Auto-retry after 5s
HTTP 4xx on gen	Expired token	Refresh NETA_TOKEN
No destinations found	Empty API response	Check token / network, retry