## User Feedback Analysis — 2026-01-26 (based on aggregated feedback through 2026-01-25) ### Data sources used - Discord daily summaries for **2026-01-23, 2026-01-24, 2026-01-25** - Recent GitHub issues/PR context referenced in the aggregation (notably migration/eligibility issue #6369; runtime/plugin fixes) --- ## 1) Pain Point Categorization (Top recurring 5–7) > Frequency notes are based on the **3 daily Discord logs** provided (Jan 23–25) plus referenced GitHub items. ### 1. Community — Tokenomics / ecosystem-token confusion & conflict (High severity, very frequent) **Observed in 3/3 days (Jan 23–25).** - Users report confusion and distrust around ecosystem token launches (e.g., **GOLD / Hyperscape funding token**), including accusations of insider behavior and “why launch coins for everything.” - Users ask for clarity on whether various tokens “impact ElizaOS or are fully independent” (question by **Alexei**, unanswered). - Mismatch between builder narrative (ecosystem growth) vs holder expectations (direct value accrual to $elizaos). **Most affected users** - Token holders/investor-aligned community members and builders worried about ecosystem reputation and attention dilution. --- ### 2. Security — Scams impersonating support & wallet-drain attempts (Critical severity, frequent) **Observed in 3/3 days (Jan 23–25).** - “Create A Ticket” bot asking for **wallet addresses**; suspicious links after posting in channel (reported by **SATA**, confirmed scam by **Odilitime**). - “Manual token migration” scam asking users to send tokens to a wallet for a promised swap (reported by **Jeburek12**). **Most affected users** - Newcomers and token holders seeking support/migration help; anyone asking questions publicly (targeted by astroturf/scam bots). --- ### 3. Technical Functionality — Upgrade/migration breakage and version conflicts (High severity, medium frequency) **Observed in 2/3 days (Jan 23–24), plus implied recurring.** - CLI upgrade from **1.6.5 → 1.7.2** causing SQL migration failures and Discord plugin bootstrap errors (reported by **YogaFlame**). - Root cause included caching and stale `package.json` references; fix required cache clearing and reinstall steps (provided by **0xbbjoker**). - Discord plugin issues: audit log permission errors; “private field access” errors fixed by upgrading to **plugin-discord 1.3.8**, but **DM issues persisted** (“User has no name or username, skipping.”). **Most affected users** - Self-hosters and developers updating existing installations; Discord operators relying on DMs/onboarding. --- ### 4. Documentation — Missing/unclear “official” guidance for key workflows (High severity, very frequent) **Observed in 3/3 days (Jan 23–25).** - External MCP registration was solved via community guidance (`POST /api/v1/mcps`) but prompted a request to “check documentation around /api/v1/mcps” (by **SATA**). - No clear “official token list / what’s official / relationship to core” documentation (Alexei’s question unanswered; broader confusion in token threads). - Migration legitimacy guidance missing (scams thrive where the official process is unclear). - DaVinci Resolve MCP work hit **API limitations**, needing clear documentation of constraints and workarounds. **Most affected users** - New builders trying to integrate MCP; token holders; anyone migrating; plugin integrators. --- ### 5. Integration — MCP adoption demand (Cloud + external tools) outpacing productized UX (Medium–high severity, medium frequency) **Observed in 2/3 days (Jan 24–25).** - Users want **external MCP servers attached to Eliza Cloud agents** (AI red teaming / HITL). - New MCP integrations emerging (DaVinci Resolve MCP; Sportradar plugin for live NBA data). - Request to “add Eliza MCP functionality to the channel” suggests missing first-class surfaces for MCP discovery/config. **Most affected users** - Power users building agents that need tool access (video editing, sports prediction, red teaming). --- ### 6. UX/UI + Product — Unclear path to “single-purpose agents” and deployment adoption (Medium severity, medium frequency) **Observed explicitly on 1/3 days (Jan 25), implicitly supported by other threads.** - “Why aren’t there more specialized one-trick-pony agents everywhere?” (**Skinny**) points to friction in packaging, sharing, hosting, or monetizing agents. - Community interest in public agent discovery exists, but users still don’t see obvious, repeatable “ship a small agent” workflows. **Most affected users** - Builders who want fast time-to-demo and distribution; newcomers deciding whether to invest. --- ### 7. Performance / Infrastructure — Confusion about practical hosting footprints and “always-on” operation (Medium severity, medium frequency) **Observed in 1–2/3 days (Jan 24–25).** - Users self-host on **mini PCs** to avoid leaving main machines on 24/7; Oracle Cloud free tier guide referenced (4 vCPU / 24GB RAM). - Indicates demand for “recommended deployments” and cost/perf guidance (what works for small agents vs production). --- ## 2) Usage Pattern Analysis (actual vs intended) ### How users are actually using elizaOS - **Tool-using agents via MCP**: external MCP servers registered to Cloud agents; early-stage “tool marketplace” behavior driven by community. - **Vertical integrations**: - **Video editing automation** (DaVinci Resolve MCP). - **Sports prediction/data agents** (Sportradar plugin). - **AI red teaming** with **human-in-the-loop** workflows. - **Crypto-adjacent community coordination**: tokens dominate attention; users evaluate ecosystem credibility as part of product adoption. - **Self-hosting as default**: mini PCs, Mac minis for local models, and free-tier cloud deployments. ### Emerging / unexpected use cases - **Creative production agents** (timeline editing, transitions, audio workflows) are a concrete “agent-as-software-worker” category. - **Prediction market / Polymarket-adjacent agents** (requested by ElizaBAO) plus live sports data integrations suggests “agents as real-time decision engines.” ### Feature requests aligned with observed usage - First-class **MCP management UX** (register, test, permissions, logs) in Eliza Cloud. - “Deployable single-purpose agent” templates and a lightweight publishing path. - More robust Discord DM/onboarding support (reliability matters for community-facing agents). --- ## 3) Implementation Opportunities (2–3 concrete solutions per major pain point) ### A) Tokenomics confusion & conflict (Community) **1) Publish a canonical “Token & Ecosystem Policy” page (High impact, low–medium effort)** - Include: - What tokens are **official**, **unofficial**, or **community-led** - Relationship to $elizaos (utility, fees, burns, none) - Disclaimers + “we will never ask you to send tokens for migration” - Similar approach: many ecosystems maintain canonical registries (e.g., “official contracts” pages and token lists). **2) Create an “Ecosystem project registry” with required metadata (High impact, medium effort)** - Each project entry: owner, audits/risks, token allocation summary, utility link to ElizaOS (if any), and status. - Similar approach: plugin registries with verification badges (common in extension marketplaces). **3) Establish communication rituals for launches (Medium impact, low effort)** - Pre-announce decision criteria, post-mortem templates for issues (e.g., liquidity mistakes), and pinned FAQs. - Similar approach: “RFC + postmortem” norms used in mature OSS communities to reduce trust shocks. --- ### B) Scams & impersonation (Security) **1) Add a prominent “Security Center” + pinned Discord banner (Critical impact, low effort)** - Pin: “Support will never DM first,” “Never share seed/private keys,” “Never send tokens to migrate.” - Include known scam patterns (wallet requests, migration addresses, ticket bots). - Similar approach: standard “anti-scam” pinned posts in major crypto/OSS Discords. **2) Harden Discord permissions & bot policy (Critical impact, medium effort)** - Audit and lock down ticket bots; require allowlisted bots only. - Add an automated moderation rule: delete messages that request wallet addresses / include known scam keywords, route to mod queue. **3) Add “Report Scam” structured flow (High impact, medium effort)** - A slash command or form that captures message link, user, screenshot, and wallet address for rapid response. --- ### C) Upgrades/migrations breaking installs (Technical Functionality) **1) Ship an “Upgrade Doctor” CLI command (High impact, medium effort)** - `eliza doctor` checks: - CLI/core/plugin version compatibility matrix - cached bun/npm artifacts - DB migration status - known bad combos (e.g., core 1.7.x with discord plugin < 1.3.8) - Similar approach: `brew doctor`, `flutter doctor` dramatically reduce support load. **2) Provide a versioned migration guide with copy/paste paths (High impact, low effort)** - A single doc page: “Upgrading 1.6.x → 1.7.x” - Include exact steps that solved YogaFlame’s issue (cache rm, reinstall global CLI, clean lockfiles). **3) Add automated preflight checks in bootstrap (Medium impact, medium effort)** - If mismatch detected, fail fast with actionable error: “Your plugin-discord is 1.3.3; required ≥ 1.3.8.” --- ### D) MCP integration demand vs limited UX (Integration + Documentation) **1) Add a Cloud UI for MCP registration/testing (High impact, medium effort)** - Wrap `POST /api/v1/mcps` with: - “Test connection” button - Example curl snippets - Request/response logging + timeouts - Similar approach: Zapier/Retool-style connector setup flows. **2) Create an “MCP cookbook” repo section (High impact, low effort)** - Include: - External MCP server template - Security model: auth headers, allowlists - Examples: DaVinci-like integration, Sportradar-like data plugin **3) Add permissions/scoping for MCP tools (Medium impact, higher effort)** - Per-agent allowlists and rate limits; avoid accidental tool abuse (important for red teaming and external endpoints). --- ### E) Adoption gap for single-purpose agents (UX/Product) **1) Publish “One Trick Pony” starter templates (High impact, low effort)** - Templates for: Discord FAQ bot, sports prediction bot, video editing assistant, red team assistant. - Each template: deploy instructions (local + Oracle free tier), minimal config, demo script. **2) Add a “Deploy & Share” flow (High impact, medium effort)** - One-click publish to a public directory (or shareable link) with required metadata. - Similar approach: Hugging Face Spaces / Langflow templates lower activation energy. **3) Clarify “intended vs best-fit” use cases on landing/docs (Medium impact, low effort)** - If the platform shines for integrations and workflows, say so; reduce mismatch expectations. --- ## 4) Communication Gaps (expectations vs reality) ### Key mismatches observed - **Expectation:** token launches are coordinated, transparent, and directly benefit $elizaos holders **Reality:** ad-hoc launches (e.g., Pump.fun “gold presale”) created reputational shocks and prolonged arguments. - **Expectation:** “support” bots can be trusted in Discord **Reality:** ticket/wallet scams are actively targeting users who ask questions. - **Expectation:** updating CLI/plugins is straightforward **Reality:** cache/version skew causes multi-step failures requiring insider knowledge. - **Expectation:** MCP is “plug and play” **Reality:** it works, but the user journey is currently API-first and doc-light. ### Recurring questions indicating onboarding/doc gaps - “Can Eliza Cloud agents use external MCP?” → answered, but indicates missing Cloud/MCP docs. - “Do these tokens affect ElizaOS or are they independent?” → unanswered, indicates missing ecosystem map. - “Why would support need my wallet?” → recurring scam pattern; needs prominent warnings. **Suggested communication fixes** - Maintain a single pinned “Start Here” post per channel: **Security**, **Upgrades**, **MCP**, **Official links**. - Add “Known Issues” section (Discord plugin DMs, upgrade pitfalls) updated weekly. --- ## 5) Community Engagement Insights ### Power users and what they need - **0xbbjoker**: repeatedly solving upgrade/plugin issues → needs a place to codify fixes (docs PRs, “known issues” ownership). - **Odilitime**: central responder for security confirmations + ecosystem clarifications → needs tooling to reduce repeated Q&A (macros, canonical pages). - **SOLOMON VANDY**: provides high-quality MCP/Cloud API guidance → good candidate to co-author MCP docs/examples. - **Irie_Rubz / PatoVeloso**: driving real MCP integrations (DaVinci) → need integration support + official “MCP cookbook.” - **LarpsAI**: hosting guidance (mini PCs / Oracle) → needs official deployment tiers and hardware guidance. ### Newcomer friction signals - Basic navigation questions (“Does Eliza Town have its own Discord?”). - Confusion during upgrades; reliance on community commands to fix caching. - Susceptibility to scams when asking for help. ### Converting passive users to contributors - Create “Docs bounty” issues for: - MCP registration guide - Upgrade doctor/migration guide - Security center / anti-scam page - Add a lightweight recognition system: weekly “Top Helper” shoutouts tied to merged docs PRs (not just Discord replies). --- ## 6) Feedback Collection Improvements ### Current channel effectiveness (from this dataset) - Discord is effective for rapid troubleshooting, but: - Answers get lost (same questions recur). - Unanswered questions persist (token relationships; adoption concerns). - Scam risk increases when users seek help publicly. ### Improvements for more structured, actionable feedback 1. **Introduce Discord “/feedback” and “/bug” forms** that auto-create GitHub issues with labels (docs/security/integration). 2. **Weekly “Top 3 pain points” poll** pinned for 24h; summarize results in GitHub Discussions. 3. **Add required metadata to reports**: version numbers, plugin list, hosting type (mini PC/Oracle/local), repro steps. ### Underrepresented segments in the current feedback - Non-crypto builders focused purely on agent workflows (feedback dominated by tokens). - Production operators (SRE-style feedback on uptime, monitoring, scaling). - Non-Discord users (no evidence from forums/email in provided data). --- ## Prioritized High-Impact Actions (next 1–2 weeks) 1. **Launch a Security Center + pinned Discord anti-scam banner** (includes “we never ask for wallets/tokens,” migration legitimacy, bot policy). 2. **Publish an “Official Tokens & Ecosystem Policy” page + canonical token/project registry link**, and answer Alexei-style questions with a single source of truth. 3. **Ship an “Upgrade Guide 1.6.x → 1.7.x” and begin an `eliza doctor` CLI spec**, using the exact fixes that resolved YogaFlame/DigitalDiva issues. 4. **Document MCP external server registration (`POST /api/v1/mcps`) with examples + add an MCP cookbook starter**, turning Solomon’s guidance into official docs. 5. **Create 3–4 single-purpose agent templates (one-trick-pony) with deploy instructions (mini PC + Oracle free tier)** to address adoption concerns raised by Skinny.