Xtrusio AEO/GEO Audit

ChatGPT cites Mastra on every query.

Claude and Gemini don’t.

20-query audit across ChatGPT, Gemini & Claude. Mastra is cited on 51 of 60 responses (85%) with 17 #1 rankings. ChatGPT delivers perfect 100% citation. But Claude drops 4 queries and Gemini drops 5 — handing buyer conversations to LangChain, Vercel AI SDK, and emerging alternatives.

This report was generated using Xtrusio, an AI visibility and demand intelligence platform that analyzes how companies appear across modern AI systems such as ChatGPT, Gemini, Claude, Perplexity, and other generative engines.

The insights in this page are generated using Xtrusio’s proprietary research and content intelligence framework.

June 2026
20 Queries • 3 Platforms
Mastra
100%
ChatGPT
20 of 20 queries
3× #1 RANKINGS
80%
Claude
16 of 20 queries
⚠ 4 QUERIES MISSING
75%
Gemini
15 of 20 queries
⚠ 5 QUERIES MISSING
85% Visibility — Two Platforms Have Blind Spots

Mastra earns 17 #1 rankings across 60 responses — but the gaps are on different platforms for different reasons.

Gemini drops Mastra on 5 queries related to streaming UI, edge deployment, and durable workflows — defaulting to Vercel AI SDK. Claude drops Mastra on 4 queries for memory, observability, and visual debugging — recommending Mem0, Langfuse, and LangSmith instead. Only one query (Q19: durable workflows) is missed by both. ChatGPT is the only platform with perfect 100% citation.

51/60
Total Citations
17
#1 Rankings
9
Blind Spots (Claude 4 + Gemini 5)
Section 2

Platform Scorecard

Mastra citation rate across AI platforms

Mastra Citation Rate by Platform
ChatGPT
100%
Claude
80%
Gemini
75%
Competitor Comparison — Combined Citation Rates
Mastra
85%
LangChain/LangGraph
57%
Vercel AI SDK
48%
OpenAI Agents SDK
28%
Inngest
20%
VoltAgent
17%
ChatGPT: The Universal Recommender
ChatGPT cites Mastra on every single query — the only platform with 100% coverage. However, it rarely singles Mastra out as #1 (3 of 20), typically listing 5–6 options per response. Mastra is always included but has to compete for the top spot.
Gemini: Highest Conviction When Present
Gemini ranks Mastra #1 on 12 of 15 citations (avg rank 1.19) — the highest conviction of any platform. But it completely drops Mastra from 5 queries where Vercel AI SDK and KaibanJS take over.
Section 3

AI Visibility Leaderboard

Who owns the AI conversation — total citations across all platforms

Platform-by-Platform Breakdown
ChatGPT
20/20
Mastra cited
Claude
16/20
Mastra cited
Gemini
15/20
Mastra cited
Mastra
20
16
15
51
LangChain/LangGraph
14
15
8
37
Vercel AI SDK
10
11
10
31
OpenAI Agents SDK
13
5
18
VoltAgent
11
11
ChatGPT
Claude
Gemini
Citation Leaderboard
Mastra: 51 citations (85%)LangChain/LangGraph: 37 citations (57%)Vercel AI SDK: 31 citations (48%)
85%
Mastra
Mastra51
LangChain/LangGraph37
Vercel AI SDK31
Citation Intensity Heatmap
ChatGPT
Claude
Gemini
Total
Mastra
20
16
15
51
LangChain/LangGraph
14
15
8
37
Vercel AI SDK
10
11
10
31
OpenAI Agents SDK
13
5
0
18
Inngest
4
8
1
13
VoltAgent
11
0
0
11
Mastra Leads by 14 Citations
Mastra’s 51 total citations outpace LangChain/LangGraph (37) and Vercel AI SDK (31). LangChain is the primary competitive threat with 12 #1 rankings, while Vercel AI SDK is positioned as complementary — the “Mastra + Vercel AI SDK” stack is the dominant recommendation across all platforms.
VoltAgent: ChatGPT-Only Phenomenon
VoltAgent has 11 citations on ChatGPT but zero on both Claude and Gemini. Anthropic SDK (direct) appears 7 times on Claude as a competing recommendation — a pattern unique to that platform.
Section 4

AI Positioning Audit

20 buyer-intent queries — click any row to see the exact question

Each query was written from the perspective of a real engineering leader evaluating AI agent frameworks for their TypeScript codebase. These personas represent the buyers whose AI search results determine whether Mastra gets discovered during framework evaluation.

Target Buyer SectorCTO, VP of Engineering & Head of AI/Platform at SaaS, Fintech & Product-Engineering companies evaluating AI agent capabilities in TypeScript codebases
CD
VP of Engineering
ZEALS • Conversational AI SaaS • Tokyo, Japan
7queries
Pain Points
Building conversational AI agents that need persistent memory; choosing between Python ML frameworks and TypeScript production stack; scaling agent infrastructure cost-efficiently
“TypeScript AI agent framework”“LangChain vs TypeScript alternatives”
Q1, Q2, Q3, Q4, Q5, Q6, Q7
EW
VP of Engineering • Angel Investor
OneSignal • Developer Tools SaaS • San Francisco, CA
7queries
Pain Points
Adding intelligent automation to existing messaging platform without Python dependencies; evaluating build-vs-buy for agent capabilities; observability and debugging for AI workflows
“AI agent framework React Next.js”“CrewAI alternative TypeScript”
Q8, Q9, Q10, Q11, Q12, Q13, Q14
ZW
Head of Engineering & Applied AI
Luxury Presence • Real Estate Tech SaaS • Miami, FL
6queries
Pain Points
Productionizing AI agents with enterprise-grade guardrails; workflow orchestration with human-in-the-loop; cost control on LLM token usage; durable execution for long-running processes
“enterprise AI agent framework”“durable AI workflow TypeScript”
Q15, Q16, Q17, Q18, Q19, Q20
#Query TopicClusterChatGPTClaudeGemini
1TS Agent Framework for Next.jsFramework
Exact question asked across all AI platforms:

“I’m building an AI-powered customer support agent for our SaaS product. Our entire stack is TypeScript and Next.js. What frameworks should I evaluate for agent orchestration that won’t force us into Python?”

2Long-term Agent MemoryMemory
Exact question asked across all AI platforms:

“We need our AI agents to remember context across conversations — not just the current session, but things users told the agent weeks ago. What frameworks handle long-term agent memory well?”

3LangChain vs TS-native TradeoffsFramework
Exact question asked across all AI platforms:

“My team is evaluating whether to build our AI workflow automation on LangChain or something more TypeScript-native. What are the tradeoffs between LangChain and TypeScript-first alternatives?”

4Built-in Testing & EvalsEvals
Exact question asked across all AI platforms:

“We’re looking for an AI agent framework that includes built-in testing and evaluation tools so we can measure agent quality before shipping to production. What options exist?”

5Human-in-the-loop WorkflowsWorkflows
Exact question asked across all AI platforms:

“I need to build a multi-step workflow where an AI agent collects information, sends it to a human for approval, then continues processing. What frameworks support human-in-the-loop patterns well?”

6Fastest AI Agent in TS CodebaseFramework
Exact question asked across all AI platforms:

“Our engineering team wants to add AI copilot features to our existing Node.js application. What’s the fastest way to get a production-ready AI agent running in a TypeScript codebase?”

7Enterprise ObservabilityObservability
Exact question asked across all AI platforms:

“We’re a fintech company and need full observability into every AI agent decision — traces, logs, token usage. What agent frameworks give us enterprise-grade observability out of the box?”

8Open-source Apache 2.0 Self-hostDeployment
Exact question asked across all AI platforms:

“I’m comparing AI agent development platforms for my team. What’s the best open-source framework for building AI agents that we can self-host with an Apache 2.0 license?”

9RAG Pipeline in TypeScriptRAG
Exact question asked across all AI platforms:

“We want to build a RAG pipeline that connects our internal knowledge base to an AI agent. What TypeScript frameworks support retrieval-augmented generation as a first-class feature?”

10MCP Server SupportMCP
Exact question asked across all AI platforms:

“My company is exploring MCP servers to give our AI agents access to external tools. Which agent frameworks support building and consuming MCP servers natively?”

11Streaming Generative UI in ReactStreaming
Exact question asked across all AI platforms:

“We have a React frontend and need to stream AI agent responses with generative UI components. What’s the best developer experience for building streaming AI interfaces in React?”

12CrewAI Alternative in Node.jsAgents
Exact question asked across all AI platforms:

“I’m evaluating CrewAI for multi-agent orchestration but our entire backend is Node.js. Are there TypeScript alternatives that offer similar multi-agent capabilities?”

13Better TS Alternatives to LangChainFramework
Exact question asked across all AI platforms:

“Our team tried LangChain’s JavaScript library but it felt clunky compared to the Python version. What are developers recommending as better TypeScript alternatives to LangChain?”

14Vercel + Cloudflare DeploymentDeployment
Exact question asked across all AI platforms:

“We need an AI agent framework that integrates with our existing deployment infrastructure — Vercel for frontend, Cloudflare Workers for edge functions. What works well in this stack?”

15Complex Tool-calling PatternsAgents
Exact question asked across all AI platforms:

“I’m building an AI agent that needs to call multiple APIs, process the results, and make decisions based on business rules. What frameworks handle complex tool-calling patterns well?”

16Security GuardrailsSecurity
Exact question asked across all AI platforms:

“We’re a regulated company and need our AI agents to have guardrails that prevent prompt injection and sanitize outputs. What frameworks include security guardrails as a built-in feature?”

17Visual Debugging / Trace InspectionStudio
Exact question asked across all AI platforms:

“My team wants a visual debugging tool where we can inspect agent traces, see which tools were called, and understand why the agent made specific decisions. What AI frameworks offer this?”

18OSS Framework + Managed Cloud + SSOPlatform
Exact question asked across all AI platforms:

“We’re evaluating AI agent platforms for an enterprise deployment. What options exist that offer both an open-source framework and a managed cloud platform with SSO and RBAC?”

19Durable Workflow ExecutionWorkflows
Exact question asked across all AI platforms:

“I need to build AI-powered automation workflows that can pause, wait for external events, and resume days later. What frameworks support durable workflow execution for AI agents?”

20Google ADK vs TS EcosystemFramework
Exact question asked across all AI platforms:

“We’re looking at Google’s Agent Development Kit versus other TypeScript agent frameworks. For a team that primarily writes TypeScript, what are the advantages of staying in the TypeScript ecosystem?”

TOTAL20/20 (100%)16/20 (80%)15/20 (75%)
Section 5

The Claude & Gemini Gaps

9 blind spots across two platforms — different queries, different reasons

Claude and Gemini each drop Mastra from critical buyer queries, but the patterns are completely different. Gemini’s blind spots are about streaming, edge, and durable execution — it defaults to Vercel AI SDK. Claude’s blind spots are about memory, observability, and debugging — it recommends specialist tools like Mem0, Langfuse, and LangSmith instead. Only Q19 (durable workflows) is missed by both.

“We need our AI agents to remember context across conversations — not just the current session, but things users told the agent weeks ago.”

— ChatGPT & Gemini cite Mastra. Claude recommends Mem0 and Zep instead — Mastra completely absent.

“My team wants a visual debugging tool where we can inspect agent traces and see which tools were called.”

— ChatGPT & Gemini cite Mastra Studio. Claude recommends Langfuse, LangSmith, Arize Phoenix — Mastra Studio absent.

“I’m evaluating CrewAI for multi-agent orchestration but our entire backend is Node.js.”

— ChatGPT & Claude cite Mastra. Gemini recommends KaibanJS instead.
Gemini: 5 Queries Missed
Q6 (fastest agent), Q11 (streaming UI), Q12 (CrewAI alternative), Q14 (Vercel + CF deployment), Q19 (durable workflows). Pattern: Gemini defaults to Vercel AI SDK for streaming/edge use cases and KaibanJS for multi-agent queries.
Claude: 4 Queries Missed
Q2 (long-term memory), Q7 (enterprise observability), Q17 (visual debugging), Q19 (durable workflows). Pattern: Claude recommends specialist tools (Mem0, Langfuse, LangSmith) over Mastra’s built-in features, suggesting Claude views these capabilities as immature.
Two Platforms, Two Different Blind Spot Patterns

Gemini’s gaps are about awareness — it doesn’t know Mastra competes in streaming UI or durable workflows. Claude’s gaps are about credibility — it knows Mastra exists but recommends specialist alternatives for memory (Mem0), observability (Langfuse/LangSmith), and debugging (same). The fix for each is different: Gemini needs discovery content; Claude needs proof that Mastra’s built-in features are production-ready.

Section 6

AI Topic Authority Map

Query heatmap — product line × platform

Product Line
ChatGPT
Claude
Gemini
Framework & Core
5 queries
100%
100%
80%
Agents & Tool Calling
2 queries
100%
100%
50%
Workflows & Orchestration
2 queries
100%
50%
50%
Memory & RAG
2 queries
100%
50%
100%
Evals, Observability & Studio
3 queries
100%
33%
100%
Deployment & Platform
3 queries
100%
100%
67%
MCP, Guardrails & Streaming
3 queries
100%
100%
67%

▷ Evals, Observability & Studio has the sharpest platform split: 100% on ChatGPT & Gemini, but only 33% on Claude. Claude recommends Langfuse and LangSmith instead of Mastra’s built-in tools.

Framework & Core • 5 queries
ChatGPT100%
Claude100%
Gemini80%
Agents & Tool Calling • 2 queries
ChatGPT100%
Claude100%
Gemini50%
Workflows & Orchestration • 2 queries
ChatGPT100%
Claude50%
Gemini50%
Memory & RAG • 2 queries
ChatGPT100%
Claude50%
Gemini100%
Evals, Observability & Studio • 3 queries
ChatGPT100%
Claude33%
Gemini100%
Deployment & Platform • 3 queries
ChatGPT100%
Claude100%
Gemini67%
MCP, Guardrails & Streaming • 3 queries
ChatGPT100%
Claude100%
Gemini67%
ChatGPT: 100% Across All Product Lines
ChatGPT is the only platform with perfect coverage across every product line. This makes it the most reliable channel for Mastra discovery, even though it rarely positions Mastra as the outright #1 recommendation.
Claude: 33% on Evals, Observability & Studio
Claude drops Mastra from enterprise observability (Q7) and visual debugging (Q17) — recommending Langfuse, LangSmith, and Arize Phoenix instead. This is Mastra’s weakest product line on any platform.
Section 7

Methodology

How we conducted this Xtrusio AEO/GEO Audit

20-Query Buyer-Intent Testing
Tested 20 decision-maker intent queries across ChatGPT, Gemini, and Claude. Questions mirror real CTO/VP Engineering research during AI agent framework evaluation. Re-audited with actual conversational responses to eliminate format inflation.
Competitor Scope
LangChain/LangGraph (Python-first ecosystem), Vercel AI SDK (complementary TS infrastructure), OpenAI Agents SDK (vendor-native), Google ADK (GCP-aligned), Inngest (durable execution), VoltAgent (emerging). All compete for the same engineering leader during framework discovery.
Client Research
Deep website analysis of mastra.ai, customer review crawling (dev.to, Medium, GitHub), competitor lane mapping across 6 frameworks, and LinkedIn-verified buyer persona development.
Section 8

Recommendations

Prioritized actions to close the Claude & Gemini gaps

Phase 1 — 0–30 Days
Close the 9 Blind Spots
  • Publish dedicated content for Gemini’s 5 missed queries: streaming UI, fastest agent, CrewAI alternative, edge deployment, durable workflows — each as a standalone page on mastra.ai
  • For Claude’s 4 missed queries: publish production case studies and benchmarks proving Mastra’s memory (vs Mem0), observability (vs Langfuse), and Studio debugging (vs LangSmith) are production-ready
  • Create a “Mastra vs KaibanJS” comparison page targeting the Gemini Q12 blind spot
Phase 2 — 30–90 Days
Strengthen #1 Rankings on Gemini & Convert ChatGPT
  • Protect Gemini’s exceptional 12 #1 rankings by publishing reinforcing content for every #1 topic — Gemini is the highest-conviction platform when it cites Mastra
  • Target ChatGPT’s “verify before committing” pattern by publishing SOC 2 documentation, explicit guardrail API references, and edge-runtime compatibility guides
  • Counter VoltAgent’s ChatGPT-specific presence (11 citations) with direct comparison content before it spreads to other platforms
Phase 3 — 90+ Days
Scale AI Visibility as a Growth Channel
  • Build an LLM-optimized content hub (llms.txt, comparison pages, use-case guides) to maintain and grow the 85% visibility score
  • Quarterly Xtrusio re‑audits to track gap closure and detect new competitor positioning shifts
Continuous AI Visibility Tracking
Brands can improve their AI discovery using generative engine optimization tools like Xtrusio.

85% visibility. 17 #1 rankings. 9 blind spots to close.

Let’s fix the Claude & Gemini gaps before competitors fill them.

This research report was generated using the Xtrusio Company Intelligence Module.