# User Feedback Analysis — 2026-02-16 (covering 2026-02-13 to 2026-02-15 signals)

## 1) Pain Point Categorization (top recurring 7)

> Quantification note: counts/percentages below are based on distinct, repeated mentions in the provided Discord summaries + GitHub issues in this bundle (≈20–25 “feedback signals” across ~15+ distinct users). Percentages are directional, not statistically rigorous.

### A. **Documentation (Highest frequency)**
**~40–45% of signals** relate to “what is this / how do I use it / what’s the roadmap” confusion.
- **Non-technical onboarding gap:** Users (e.g., *Rainman, yojo*) asked how a “non-tech guy” can benefit; answer implied real value needs plugins + technical setup.
- **Roadmap not readable for non-coders:** Requests for milestone clarity (zero-code timeline, token utility/governance, payment via token without wallet connect).
- **Unanswered “what is Milaidy / does it tie to $elizaos?”** (user *g*) suggests missing “ecosystem map” docs.

### B. **Community / Trust & Safety (High severity)**
**~15–20% of signals**, but severe.
- **Active scam attempts in Discord threads** (reported scammer ID `364718078946312192`; fake “support ticket” scam targeting *fragmtagmbagm*).
- **Channel/thread permissions enabling scams:** explicit suggestion to restrict thread creation.

### C. **Integration (High impact for real users building agents)**
**~15–20% of signals**.
- **Need custom OpenAI-compatible endpoints** (GitHub #6490) for providers like SiliconFlow.
- **“Base install should prompt ElizaCloud capabilities”** (DorianD) indicates discoverability/integration friction between local framework and cloud features.
- **Cross-platform agent deployment help** (Spartan agent deployment request; Moltbook anti-bot flow).

### D. **Technical Functionality (Product bugs / correctness)**
**~15–20% of signals**.
- **Duplicate LLM calls when messages contain URLs** in webapp (GitHub #6486): doubles cost + duplicates output.
- **Image content stripped in cloud chat** (GitHub #6494): multimodal reliability issue.
- **Agent posting issues on X (roomId-related error)** (reported by *Gamer*; unresolved in thread).

### E. **UX/UI (Expectation management + in-product guidance)**
**~10–15% of signals**, overlaps with docs but shows up as “journey” problems.
- Users can register on **elizacloud.ai** but don’t see safe/recommended plugins or “next steps” for real use cases.
- Migration/support flows: users looked for tickets after closure; UX didn’t prevent late arrivals from wasting time.

### F. **Community / Token Administration (Very high emotion + reputational risk)**
**~20–25% of Discord discussion** was dominated by this.
- **Token migration deadline distress** (e.g., *fibsonly, Andi CEGY, GomesCozendey, crunchy_vertex*): long-term holders missed a 90-day manual ticket window; post-deadline support locked.
- Users argued on-chain proof should allow manual exception; moderators confirmed “nothing can be done.”

### G. **Product Strategy / Tokenomics (Strategic pain point)**
**~10–15% of signals**, but critical to community narrative.
- “Framework has no direct tie to token → zero reason to buy” (*DannyNOR NoFapArc*).
- Questions about low market cap vs meme coins; nostalgia about lost memetic power moving away from ai16z branding.
- Mentioned idea: **buybacks using Eliza Cloud revenue** (*Alexei*), staking “after JEJU” without a timeline.

---

## 2) Usage Pattern Analysis (actual vs intended)

### How users are actually using elizaOS
1. **Builders integrating agents into hostile/real environments**  
   Example: *funboy* built a Moltbook anti-bot verification solver (obfuscated math, strict penalties), intercepting `verification_required` and posting answers back within 30s.  
   **Pattern:** elizaOS is being used as an automation/integration runtime, not just “chat with an agent.”

2. **Cloud as the entry point for non-technical users—but they hit a capability cliff**  
   *yojo/Rainman* discussion shows non-devs want “life improvement” and investor-ready demos, but current guidance is “you need plugins + technical work.”

3. **Agents as social/content operators (X/Twitter) + media pipelines**  
   *Gamer* posting issue; *0xqueen* offering seedance2.0 media account for promo videos.

4. **Community strongly uses Discord for support—despite it being scam-prone**  
   Migration seekers are funneled into Discord; scammers exploit that, especially in threads.

### Emerging / unexpected use cases
- **PR evaluation + incentivization tooling** (TipCat/tip.md demo): desire to reward documentation/explanations, not just code.
- **Multi-language expansion interest** (weekly summary: Rust/Python support) suggests more “systems/programming” users than pure JS-only audience.

### Feature requests aligned to observed usage
- Custom OpenAI endpoint (OpenAI-compatible providers) (#6490) aligns with “swap inference backends frequently.”
- Fix URL double-processing (#6486) aligns with “agents that browse/link-share frequently.”
- ElizaCloud awareness prompting in base install aligns with “users don’t discover the cloud path organically.”

---

## 3) Implementation Opportunities (2–3 concrete solutions per major pain point)

### 1) Migration deadline fallout (Category: Community + UX/UI + Documentation)
**Impact:** Very high (trust/reputation), **Difficulty:** Medium (policy + tooling).
1. **Publish a “Migration Postmortem + Final State” page** (high impact, low difficulty)  
   - Single canonical URL: dates, ratio (1:6), deadline, why no extension, what users can/can’t do now, scam warnings, and how to verify official comms.
   - Similar approach: major projects run public postmortems after contentious events to reduce repeated support load.

2. **If policy allows: one-time “proof-based claims” process with strict controls** (very high impact, medium/high difficulty)  
   - Time-boxed (e.g., 14 days), require on-chain proof + signed message from snapshot wallet + KYC-light friction only if necessary.  
   - Comparable pattern: **Merkle-claim style** airdrops/claims with explicit windows + cryptographic proofs; reduces “manual ticket” ambiguity.

3. **In-product deadline banners (Cloud + Discord + GitHub pinned)** (high impact, low difficulty)  
   - Banner on elizacloud.ai dashboard and Discord announcement pins during critical windows; include “add to calendar” links.

---

### 2) Discord scams via threads (Category: Community / Security)
**Impact:** Very high, **Difficulty:** Low/Medium.
1. **Restrict thread creation + lock sensitive channels** (high impact, low difficulty)  
   - Implement role-gated thread creation in 💬-coders / support-adjacent channels.
   - Immediate action: ban reported scammer ID `364718078946312192`, close compromised threads (already proposed).

2. **Create an official “Support Verification” bot message + auto-DM** (high impact, medium difficulty)  
   - When keywords appear (“ticket”, “migration”, “support”), bot replies: official support never asks to join other servers/Zoom; list official links only.

3. **Move support to an auditable system** (medium impact, medium difficulty)  
   - GitHub Discussions or a ticketing form with public “known issues” reduces reliance on ephemeral DMs/threads.
   - Similar projects often funnel support into tracked issues/discussions to reduce impersonation risk.

---

### 3) Non-technical onboarding + roadmap clarity (Category: Documentation + UX/UI)
**Impact:** High, **Difficulty:** Medium.
1. **Add a “Choose your path” onboarding page + Cloud dashboard card** (high impact, low/medium difficulty)  
   - Paths: (a) Non-technical (Cloud templates), (b) Builder (CLI + plugins), (c) Integrator (MCP/A2A), (d) Investor/partner (capabilities + milestones).
   - Include “what you can do today” vs “planned.”

2. **Roadmap reformat for non-coders** (high impact, medium difficulty)  
   - Convert GitHub roadmap into a simple milestone table: feature, user benefit, ETA confidence (High/Med/Low), dependencies.  
   - Address recurring asks: zero-code timeline, token utility/governance, JEJU/staking timing.

3. **Curated “safe/recommended plugins” list** (high impact, medium difficulty)  
   - In Cloud: show vetted plugins with badges (Maintained, Audited, Popular, Cost profile).  
   - Similar to curated marketplaces (e.g., “verified” integrations) to reduce choice paralysis and risk.

---

### 4) Core product bugs causing cost/quality regressions (Category: Technical Functionality + Performance)
**Impact:** High (cost + trust), **Difficulty:** Medium.
1. **Fix URL double-processing in webapp (#6486)** (high impact, medium difficulty)  
   - Add dedupe logic before LLM invocation: treat URL preview metadata as enrichment, not a second message.  
   - Add regression test: “message with URL triggers exactly 1 completion.”

2. **Fix multimodal payload loss in cloud chat (#6494)** (high impact, medium difficulty)  
   - Add contract tests for `/api/v1/chat/completions` ensuring images survive `convertToUIMessages`.  
   - Add logging/telemetry: count messages with image parts vs delivered-to-provider parts.

3. **Cost guardrails + duplication detection** (medium impact, medium difficulty)  
   - Leverage/extend scenario cost evaluation patterns (see closed issue #5759) to flag “2x calls per user message” anomalies in staging.

---

### 5) Integration flexibility: OpenAI-compatible endpoints (Category: Integration)
**Impact:** High for builders, **Difficulty:** Low.
1. **Add `OPENAI_BASE_URL` (or provider option) to OpenAI provider (#6490)** (high impact, low difficulty)  
   - Standard pattern used across SDKs: base URL override + headers override.
2. **Document “OpenAI-compatible providers” setup** (medium impact, low difficulty)  
   - Include examples: SiliconFlow, local gateways, reverse proxies.
3. **Add a provider conformance test** (medium impact, medium difficulty)  
   - Simple test suite validating chat completions, streaming, tool calls against a mocked OpenAI-compatible endpoint.

---

### 6) ElizaCloud discoverability from base install (Category: Integration + UX/UI)
**Impact:** Medium/High, **Difficulty:** Low/Medium.
1. **Add base system prompt / bootstrap hinting** (high impact, low difficulty)  
   - As proposed by *DorianD*: agents “know” ElizaCloud is an option for storage, deployment, capabilities—especially in forks like Milaidy.
2. **CLI “Connect to Cloud” step** (medium impact, medium difficulty)  
   - After `elizaos create`, prompt: “Do you want hosted memory/storage, keys, deploy?” with one command.
3. **Telemetry-free, local-only “capabilities checklist”** (medium impact, low difficulty)  
   - Print detected capabilities and recommended next integrations.

---

### 7) Token utility narrative mismatch (Category: Community + Documentation + Product Strategy)
**Impact:** High (sentiment), **Difficulty:** High (requires product decisions).
1. **Publish a “Token Utility: current vs planned” truth table** (high impact, low difficulty)  
   - Explicitly state what the token does today (if anything) and what is exploratory.
2. **Tie token to measurable ecosystem actions (when ready)** (high impact, high difficulty)  
   - Examples: discounts/credits, staking for resource access, governance for registry curation—only if aligned with legal/operational constraints.
3. **If buybacks are planned, publish mechanism constraints** (medium impact, medium difficulty)  
   - Avoid vague hints; specify triggers, transparency, and what revenue sources qualify.

---

## 4) Communication Gaps (expectations vs reality)

1. **“Zero-code agents for real use cases” expectation vs current plugin reality**  
   - Recurring questions imply users expect ChatGPT-like immediate usefulness. Current answer: without plugins/technical work, differentiation is limited.

2. **Staking/JEJU timing ambiguity**  
   - “After launching JEJU” was stated, but no schedule (asked by *Rainman* and others). This creates repeated questions and speculation loops.

3. **Migration support discoverability**  
   - Users kept asking how to open tickets after closure (*GomesCozendey*). The system didn’t intercept late users with a clear “closed” explanation + safe next steps.

4. **Milaidy relationship to ElizaOS and $elizaos is unclear**  
   - Unanswered question indicates missing official explainer: what is a fork/app, what is official vs community-built, and whether tokens are affiliated.

**Specific improvements**
- Add a pinned Discord “Start here (Non-tech / Dev / Token)” post with links.
- Add an FAQ entry: “Are Milaidy tokens official?” “What is JEJU?” “Where are official support channels?”
- Add “status pages” for: migration (closed), staking (planned), JEJU (in progress), cloud (live).

---

## 5) Community Engagement Insights

### Power users observed (and their needs)
- **Odilitime**: frontline security + support + technical clarifications; needs better anti-scam tooling, a structured support queue, and clearer “official” comms to point to.
- **yojo**: articulates product positioning and onboarding gaps; needs a non-technical narrative + roadmap artifacts.
- **DorianD**: ecosystem strategy; needs product hooks (base install prompts, cloud integration) and investor communication channels (email updates).
- **funboy**: high-skill integrator solving real-world constraints; needs better integration docs/patterns (intercept flows, challenge-response patterns).
- **Stan ⚡ / 0xConsole**: implementation contributors; benefit from smaller, well-scoped PRs and clearer contribution standards (also echoed indirectly by the “scope creep” review commentary in PR discussions).

### Common newcomer questions indicating friction
- “Am I late for migration?” / “How do I open a ticket?” (repeated across multiple users)
- “How do non-tech users benefit today?”
- “What is Milaidy / does it tie to the token?”
- “When is JEJU / staking?”

### Converting passive users into contributors
- Create “good first issue” tracks tied to pain points: docs (FAQ/roadmap visuals), integration tests, Discord moderation automation.
- Formalize a **Community Safety Team** role with clear permissions (thread management, scam escalation).
- Use TipCat-style incentives to reward **documentation and explanation PRs**, not only code.

---

## 6) Feedback Collection Improvements

### Current channel effectiveness
- **Discord:** fast, high volume, but unstructured and vulnerable to impersonation/scams; hard to convert into actionable tickets.
- **GitHub issues:** structured and actionable (e.g., #6486, #6490, #6494), but many end-users never file there.

### Improvements to gather structured, actionable feedback
1. **Add a “Report a bug / request a feature” form that routes to GitHub with templates**  
   - Pre-fills environment, repro steps, expected/actual, logs.
2. **Weekly “Top 5 known issues” post pinned in Discord + linked to GitHub**  
   - Reduces repeated questions and scam exposure.
3. **Tag feedback by persona (Non-tech / Builder / Operator / Investor)**  
   - Quick emoji reaction or dropdown in forms; helps detect under-served segments.

### Underrepresented segments
- **Enterprise/production operators** (SRE-style feedback): little direct feedback beyond security/auth summaries.
- **Non-English users/timezone-separated holders** affected by migration comms: implied by “I was just holding” and missed deadline; not enough insight on how they consume updates.
- **Non-technical cloud-only users**: present in questions, but lacking structured feedback about where they drop off in onboarding.

---

## Prioritized High-Impact Actions (next 2–4 weeks)

1. **Lock down Discord scam vectors immediately**: ban reported scammer ID, restrict thread creation, add automated “official support never DMs/Zoom/other servers” warnings on trigger keywords.  
2. **Ship fixes for the highest-cost product regressions**: resolve URL double-processing (#6486) and cloud image stripping (#6494) with regression tests.  
3. **Publish a single canonical “Migration is closed” + “JEJU/staking status” + “Milaidy explained” page and pin it everywhere** (Discord + Cloud dashboard).  
4. **Implement OpenAI provider `base_url` override (#6490) + doc page for OpenAI-compatible providers** to unblock common integration patterns.  
5. **Create a non-technical onboarding path in Cloud**: “recommended plugins,” “what you can do today,” and a roadmap snapshot with confidence levels.