## User Feedback Analysis — 2026-02-18 (based on Feb 15–17 sources)

### 1) Pain Point Categorization (Top recurring friction areas)

> **Method note (quantification):** Counts below are based on distinct, repeated mentions across the available Feb 15–17 Discord summaries + surfaced GitHub issues/PR context in the provided dataset. With a small sample, percentages are directional.

#### A. **Documentation** (high frequency, high severity)
1) **Spartan setup is under-documented and brittle** (≈29% of captured pain-point mentions)
- Users report `bun install` hanging and needing to **manually clone multiple plugins** (e.g., `@elizaos/plugin-evm`, `plugin-farcaster`, `plugin-jupiter`, `plugin-knowledge`, `plugin-mysql`, `plugin-solana`).  
- Docker is expected to work but **Docker files are currently non-functional**, worsening onboarding.
- Example: Einav Livne’s setup required manual cloning; Odilitime acknowledged Spartan “hasn’t been polished yet.”

2) **Unclear “required vs optional” plugins & installation expectations**
- Users don’t know what must be present to get a minimal agent running vs a full stack.

**Most affected:** newcomers and builders trying to deploy Spartan or run agents locally.

---

#### B. **Technical Functionality** (high frequency, high severity)
3) **Package distribution / CLI install confusion (milady)** (≈12%)
- `npx milady` installs an **unrelated Alibaba tool**, not the intended pre-release project.
- Workaround is “use GitHub release binaries,” but that’s not discoverable.
- Example: Issue opened (#324) after user davidhq hit the problem.

4) **Duplicate LLM calls when sending URLs (webapp)** (GitHub issue #6486) (≈6% but high severity impact)
- URL messages processed twice (text + attachment preview) causing:
  - **2x token cost**
  - duplicated streamed output (bad UX)
- This is a cost + trust issue (users perceive “the agent is bugging out” and/or “billing is wrong”).

---

#### C. **Integration** (medium frequency, high severity)
5) **Solana / website integration path unclear** (≈12%)
- Users asked whether ElizaOS Solana plugin is usable, what “website integration” looks like, and whether `mcp-gateway` is the right repo—questions remained unanswered.
- Example: Clawlana-Kalshi raised integration questions for Kalshi prediction market flows.

6) **OpenAI-compatible provider support incomplete (custom endpoint URL)** (GitHub issue #6490) (≈6%)
- Users want to point the OpenAI provider at third-party OpenAI-compatible endpoints (e.g., SiliconFlow), but can’t.

---

#### D. **Community** (medium frequency, high severity)
7) **Scams + support impersonation in Discord threads** (≈12%)
- Reports of scammers operating via “coders” thread and fake support tickets; users were targeted after days of interaction.
- Requested mitigation: restrict thread creation; ban specific user ID(s).
- Example: Odilitime emphasized legit support won’t move users to other Discords/Zoom.

---

#### E. **UX/UI + Communication** (high frequency, medium severity)
8) **Token-related expectations mismatch (utility, buybacks transparency, migration finality)** (≈24%)
- Repeated questions: “What’s the use of $elizaos now?”, “When buyback and will it be public?”, plus distress over **missed ai16z→ElizaOS migration deadline**.
- Example: multiple long-term holders requested exception pathways; staff repeatedly said migration is over and warned about scams.
- Users also asked why projects “don’t fall under elizaos” (brand/structure confusion).

---

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

**Observed actual usage patterns**
- **Agent builders are assembling a “plugin stack” manually** (Spartan + multiple plugins) rather than using a smooth “one command” bootstrap. The reality is closer to “monorepo-style dev environment assembly” than a product-like install.
- **Crypto/trading remains a primary use case**:
  - Security Oracle API for trading agents (RugCheck/GoPlus/sentiment, Sybil farm detection, insider concentration).
  - Kalshi prediction markets plugin work via Solana execution.
  - Trust/discovery layers (MoltBridge) motivated by “malicious skills bypassing vetting.”
- **Media/news automation is emerging as a flagship showcase use case**:
  - elizaos.news: automated daily news video generation (PlayCanvas rendering + subtitles + narration).
  - Upcoming “1–1 interview → generated show” is a creator workflow, not just dev tooling.

**Unexpected / emerging use cases**
- **Agents participating autonomously in community spaces** (Dawn in Discord) raised governance questions: “Are autonomous agents allowed here?” This suggests an emerging “agents-as-community-members” mode that needs explicit policy + tooling.

**Feature requests aligning with usage**
- “Base install should include prompts making agents aware of ElizaCloud capabilities” aligns with observed confusion about what ships in core vs cloud vs forks.
- Custom OpenAI endpoint aligns with real-world usage: many teams run OpenAI-compatible inference providers.

---

### 3) Implementation Opportunities (solutions per major pain point)

#### Pain Point #1: Spartan setup friction (Docs + Technical)
**Proposed solutions (prioritized by impact/difficulty)**
1) **Ship a verified “Spartan Quickstart” that works end-to-end** (High impact / Medium effort)
- A single canonical path: `git clone → bun install → bun dev` with a minimal agent.
- Include an explicit **plugin matrix**: required/optional + compatibility versions.
- Add a “known-good” lockfile/commit SHA references for a stable setup.

2) **Make Docker real (or remove it until it is)** (High impact / Medium effort)
- Provide a minimal `docker compose up` that starts:
  - core runtime
  - one database option (sqlite/postgres)
  - one example agent
- If Docker is not supported, explicitly label it “WIP” and hide from primary docs to reduce expectation debt.

3) **Add a preflight installer / doctor command** (High impact / Higher effort)
- `elizaos doctor` checks:
  - missing plugins
  - bun/node versions
  - environment variables
  - resolves “bun install hang” common causes
- **Comparable approach:** Next.js `create-next-app` + Prisma “preflight checks”; Homebrew’s `brew doctor`.

---

#### Pain Point #2: Package distribution confusion (`npx milady`) (Technical Functionality)
1) **Claim the npm package name (or rename clearly) + add install guards** (High impact / Medium effort)
- Publish a protected placeholder immediately if needed (even if pre-release), with clear messaging and a redirect in README.

2) **Make `npx` fail loudly with a helpful message until ready** (Medium impact / Low effort)
- If package cannot be safely published, publish a tiny package that prints:
  - “milady is pre-release; download binaries here…”
  - exact links + checksum guidance

3) **Add a “distribution status” badge across repos** (Medium impact / Low effort)
- “npm: not available / pre-release” avoids repeated user confusion.
- **Comparable approach:** HashiCorp tools clearly separate “install via releases” vs package managers until stable.

---

#### Pain Point #3: Token migration + token utility confusion (Communication + Community)
1) **Create a single canonical “Token Status” page updated weekly** (High impact / Low effort)
- Migration status: closed date, rationale, scam warnings, what support can/can’t do.
- Buyback policy: what “buybacks using revenue” means, disclosure cadence, where to verify.

2) **Pin an “Investor/Token FAQ” in Discord + website** (High impact / Low effort)
- Repeated questions indicate users are searching chat history instead of finding authoritative answers.

3) **Separate “framework” vs “token” narratives in onboarding** (Medium impact / Medium effort)
- A short diagram: elizaOS (OSS framework), Eliza Cloud (product), tokens (economic layer).  
- **Comparable approach:** Arbitrum / Optimism docs separate governance token utility from developer platform docs to prevent mismatched expectations.

---

#### Pain Point #4: Scam activity + support impersonation (Community + UX)
1) **Lock down thread creation + add “Verified Support” workflow** (High impact / Medium effort)
- Restrict thread creation to trusted roles; require approval for “support” labeled threads.

2) **Automated scam heuristics + warning banners** (Medium impact / Medium effort)
- Bot posts reminders when keywords appear: “ticket”, “support”, “DM me”, “Zoom”, “other server”.
- **Comparable approach:** Many crypto Discords use AutoMod + dedicated “Support is never in DMs” banners.

3) **Incident reporting template** (Medium impact / Low effort)
- A form: attacker ID, channel, screenshots, transaction links—improves response speed and evidence quality.

---

#### Pain Point #5: Integration ambiguity (Solana/web, identity standards, OpenAI endpoints) (Integration + Docs)
1) **Publish “Integration Paths” docs** (High impact / Low effort)
- “Web app integration” options (SSE vs websockets, MCP gateway, embedding patterns).
- “Solana plugin status” with versioned examples.

2) **Support custom OpenAI base URL** (High impact / Medium effort)
- Add `OPENAI_BASE_URL` (or provider config) and ensure it works for OpenAI-compatible vendors.
- **Comparable approach:** OpenAI SDKs + LiteLLM commonly support base URL overrides to unify compatibility.

3) **Identity/trust integration RFC** (Medium impact / Medium effort)
- Standardize how MoltBridge-like signing plugs into ElizaOS (request signing, key management, trust graph ingestion).
- Provide a small reference implementation (TS first).

---

### 4) Communication Gaps (expectations vs reality)

1) **“Docker exists” implies “Docker works”**
- Reality: Docker files are non-functional; users waste time expecting a supported path.

2) **“Pre-release” isn’t communicated at the point of install**
- `npx milady` failure shows users discover pre-release status only after hitting a broken path.

3) **Token/ops questions get deferred to “ask ops later”**
- Users want a stable reference source; repeated Qs indicate answers are not being captured in docs.

4) **Unanswered integration questions stall adoption**
- Example: Kalshi-Solana + `mcp-gateway` questions remained unanswered, signaling unclear ownership for integration support.

**Suggested improvements**
- Add “Supported / Experimental / Deprecated” labels across:
  - Spartan install methods (bun/manual/Docker)
  - plugins (Solana/EVM)
  - distribution channels (npm/releases)
- Create a “Who to ask” routing guide (partnerships, integrations, security, token).

---

### 5) Community Engagement Insights

**Power users / key contributors observed**
- **Odilitime**: primary responder for setup, token questions, security warnings; also driving major refactors.
- **Jin**: shipping elizaos.news automation; building new interview feature.
- **Dawn / Vlt9**: pushing identity/trust + security oracle architecture; seeking founding integrators.
- **Kenk / Omid Sa**: authoritative answers on migration finality; scam warnings.

**Newcomer friction signals (recurring questions)**
- “How do I deploy Spartan?” / “Why does bun install hang?”  
- “Are you recruiting?” / “Who handles partnerships?”  
- “Am I late for migration?” / “Can I open a ticket?”  
These indicate onboarding is still Discord-driven rather than doc-driven.

**Converting passive users into contributors**
- Turn Spartan setup pain into a contribution pathway:
  - “Docs bounty: reproduce Spartan install from scratch and PR improvements.”
- Create “Integration Office Hours” (30 min weekly):
  - Solana plugin, MCP gateway, OpenAI-compatible endpoints
- Offer a “Founding Agents” cohort (matches MoltBridge’s ask for 50 founding agents):
  - clear checklist, badge/role, public list of integrators, lightweight rewards.

---

### 6) Feedback Collection Improvements

**Current channel effectiveness**
- **Discord:** high velocity, good for discovery, but issues get lost; unanswered questions persist.
- **GitHub issues:** actionable for bugs/features (e.g., URL double-call, custom endpoint), but many users never file issues.

**Improvements**
1) **Add a structured “Bug report” slash command in Discord** → auto-creates a GitHub issue draft
- Captures: version, logs, repro steps, plugin list.

2) **Weekly “Top 5 recurring questions” digest** pinned + mirrored to docs
- Reduces repeat load on core responders.

3) **Underrepresented segments to actively recruit feedback from**
- Non-crypto builders (SaaS/enterprise agents) are not visible in this dataset.
- Windows users / lower-devops-skill users (likely most affected by install issues) are under-sampled.
- Create a short onboarding survey in CLI (`elizaos create`) to capture OS, intent, blockers.

---

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

1) **Publish a Spartan Quickstart + plugin requirement matrix + “Docker status” clarity** (reduce install churn immediately).  
2) **Fix/mitigate milady distribution confusion**: publish a safe npm placeholder or rename + make `npx` behavior deterministic.  
3) **Ship fixes for cost-amplifying bugs**: prioritize GitHub #6486 (duplicate URL LLM calls) due to direct cost + UX damage.  
4) **Create a single canonical Token/Migration/Buyback FAQ page** and pin it in Discord; update weekly.  
5) **Harden Discord against scams**: restrict thread creation + verified support messaging + lightweight reporting workflow.