## Intel — 2026-01-29 (ElizaOS)

### 1) Data Pattern Recognition (Velocity, Trends, Adoption, Pain-Point Correlation)

**Development velocity (Jan to date, elizaos/eliza)**
- PRs: **38 opened / 22 merged** (merge rate **58%**)
- Issues: **93 opened / 52 closed** (net +41 backlog)
- Contributors: **31 active**
- Code churn: **+25,706 / -8,753 across 269 files**; **379 commits**
- Trend signal: high throughput but rising backlog; multiple parallel tracks (v2 runtime, workflow automation, onboarding/docs, infra analytics) increases dependency risk.

**Adoption / usage signals (from Discord)**
- **v2.0.0 “next”** is actively being installed/tested (`@elizaos/core@next`), with workflow automation (n8n) positioned as the primary path for “95%” of users.
- **Socket.io** emerging as the practical alternative to **SSE** for real deployments (workaround adopted by at least one developer after MIME mismatch failures).
- **Developer tooling**: clear negative outcomes from Cursor **AUTO** mode (breakage after long troubleshooting) and fast recovery via **Composer 1**.

**Pain-point correlations across channels**
- **Embeddings instability + configuration confusion**: OpenRouter embeddings repeatedly reported as failing; OpenAI works reliably. Confusion compounded by enum validation errors (`EMBEDDING_PROVIDER`), plus cost blowups when contextual embeddings are enabled.
- **Streaming transport issues**: SSE “MIME_TYPE_MISMATCH” correlates with backend deployment/config mismatches; users resolve by switching transport rather than fixing root cause.
- **Onboarding friction**: recurring need for “how do I Git / how do I PR” guidance; strong push internally to treat docs/knowledge as “agent prompt surface” and automate bootstrapping.

---

### 2) User Experience Intelligence (Impact, Themes, Opportunities, Sentiment)

#### A. High-impact user problems (blockers)
1) **Migration portal not detecting ai16z tokens (Phantom)**
- Multiple reports; creates direct user loss risk (missed deadline, investor distrust).
- **Opportunity**: add deterministic “manual address input + snapshot proof view” and clearer error states; publish status + ETA.

2) **Knowledge plugin embeddings reliability + cost**
- Reliability: OpenRouter embeddings frequently error; OpenAI “only one that works” sentiment.
- Cost: contextual embeddings (`CTX_KNOWLEDGE_ENABLED=true`) can spike costs dramatically for small docs.
- **Opportunity**: “safe default” config preset + automatic warnings when CTX_KNOWLEDGE is enabled; provider health-check + fallback strategy.

3) **SSE streaming setup failures**
- MIME mismatch errors; workaround = socket.io.
- **Opportunity**: make socket.io the “recommended default” in docs/templates; add deploy-time validation for correct content-type and reverse proxy config.

#### B. Medium-impact friction (retention / productivity)
- **Docs discoverability**: explicit request to move “Quick Start” higher in `llms.txt`; recurring onboarding Q&A implies docs not surfacing at point-of-need.
- **Tooling pitfalls**: Cursor AUTO mode causing long, expensive dead-ends; need an “officially recommended dev loop”.

#### C. Sentiment & community health
- Mixed: strong technical momentum (MCP merged, n8n direction, plugin work) but community expresses:
  - **Quiet Discord** perception (lower ambient engagement).
  - **Token value proposition questions** + migration confusion.
  - Desire for **more “soul”** in @elizaos presence (brand/personality gap).

---

### 3) Strategic Prioritization (Impact vs Risk, Dependencies, Resource Allocation)

#### Priority 0 — Stop the bleeding (7-day horizon)
1) **Migration portal detection reliability**
- **User impact**: very high (funds + trust).
- **Tech risk**: medium (wallet/provider variance, snapshot indexing).
- **Actions**
  - Add “address-based lookup” path independent of wallet connector.
  - Publish a short status post + troubleshooting flow (Phantom-specific).
  - Instrument failures: connector errors vs snapshot-miss vs RPC issues.

2) **Embeddings: ship a “known-good” baseline + guardrails**
- **User impact**: high (cost + core feature reliability).
- **Tech risk**: low-medium.
- **Actions**
  - Provide 2 presets: `KNOWLEDGE_MIN_COST` (CTX off) and `KNOWLEDGE_MAX_RECALL` (CTX on).
  - Add runtime warning: “CTX enabled; expected cost multiplier” and show estimated chunk/token counts.
  - Run a small compatibility matrix test for Ollama/OpenRouter/OpenAI embeddings and publish results.

3) **Make socket.io the default streaming path (until SSE is hardened)**
- **User impact**: high (setup success).
- **Tech risk**: low.
- **Actions**
  - Update templates + docs: “Use socket.io unless you have a reason for SSE”.
  - Add a deployment checklist: reverse proxy headers, MIME types, route pathing.

#### Priority 1 — Unlock the platform bet (14–30 day horizon)
4) **n8n automation integration (core v2 differentiator)**
- Current: plugin-n8n-workflow reported **~30% complete**; architecture defined (Generator → Engine/cron → OAuth gateway).
- **Critical dependencies**
  - **OAuth gateway** (multi-tenant credentials) is the gating item for most real workflows.
  - Workflow validation/testing harness (credential requirements + “experimental” workflow lifecycle).
- **Resource allocation recommendation**
  - Assign a single “workflow platform owner” to drive end-to-end: credentials → node discovery → execution → debugging loop.
  - Timebox: ship **10 prebuilt workflows** (gmail/calendar/notion/linear basics) before expanding breadth.

5) **Onboarding bootstrapping skill + self-maintaining knowledge config**
- MCP merged + dynamic repo/channel tracking is a leverage point.
- **Actions**
  - Ship “Bootstrap Skill v1”: connects agent to docs + GitHub activity + Discord knowledge, with a “self-diagnose setup” command.
  - Implement monthly automated config review (active/inactive repos/channels) to prevent knowledge rot.

#### Priority 2 — Cost/performance optimization & product surface
6) **Intelligent model routing (small local router)**
- **User impact**: medium-high (cost control, latency).
- **Tech risk**: medium (routing errors degrade UX).
- **Recommendation**
  - Start with heuristic routing (tokens/context size + action type) before training a local model.
  - Add observability: route choice, cost, latency, user satisfaction proxy.

7) **@elizaos “personality” agent**
- **User impact**: medium (top-of-funnel, engagement).
- **Risk**: medium (brand safety, compliance).
- **Recommendation**
  - Treat as a product experiment with guardrails: content policy, escalation, and “approved topics” list.
  - Tie to measurable goals: follower growth, click-through to landing → poke → conversation.

---

### Operational Recommendations (Actionable, Quantifiable)

**A. Reduce onboarding time-to-value**
- Add an “Official Dev Setup” page: **Composer 1 recommended**, Cursor AUTO discouraged, known failure modes.
- Move Quick Start higher in `llms.txt`; add “First 30 minutes” checklist.
- KPI: reduce “setup help” questions in Discord by **30%** over 2 weeks.

**B. Reliability hardening**
- Publish a “Transport decision” doc: SSE vs socket.io vs WebSocket, with default = socket.io.
- Add preflight checks in starter templates (validate content-type / route availability).
- KPI: cut streaming-related support threads by **50%**.

**C. Knowledge plugin stabilization**
- Add runtime warnings for CTX cost; provide embedding provider health-check.
- KPI: reduce embedding-provider error reports to near-zero; track average $/query and p95 latency.

**D. Community trust & governance**
- Provide a clear statement on legal/liability posture for agent-operated dapps (esp. gambling-like mechanics) and what “fully decentralized” means operationally.
- Publish migration + airdrop communications in one canonical place; link from Discord bots.
- KPI: fewer repeated migration questions; improved sentiment in discussion channel.

---

### Watchlist (Risks & Leading Indicators)
- **Backlog growth** (+41 issues net this month) suggests scope pressure; risk of shipping half-integrations.
- **n8n + OAuth** is the main dependency chain; slipping OAuth blocks the “95% workflow” thesis.
- **Embeddings provider inconsistency** undermines knowledge experience; if not fixed, it will force OpenAI-only perception.
- **Migration failures** + unclear token value proposition: ongoing reputational drag unless proactively addressed.