Key takeaways
- Raised a $22M Series A led by Spark Capital in April 2026 ($35M total) and launched the Mastra platform: Studio, Server, and Memory Gateway
- v1.0 stable shipped January 2026; ~25k GitHub stars and 90+ model providers as of June 2026
- Observational Memory's 94.87% on LongMemEval (gpt-5-mini) remains the highest published score Mastra cites as of June 2026
FAQ
What is Mastra?
Mastra is a TypeScript framework for building AI agents and applications, featuring agents, workflows, RAG, and a novel memory system.
Who created Mastra?
The founders of Gatsby — Sam Bhagwat, Abhi Aiyer, and Shane Thomas — backed by Y Combinator W25, with a $13M seed and a $22M Series A led by Spark Capital. Tyler Barnes (ex-Netlify, ex-Gatsby) built Observational Memory.
What is Observational Memory?
A memory system that uses background agents to compress conversations into dense observations, achieving SOTA benchmark scores without RAG or graphs.
How does Mastra compare to LangChain?
Mastra is TypeScript-native and batteries-included; LangChain is Python-first with more modularity but higher integration complexity.
Is Mastra open source?
Yes, the framework is open source under Apache 2.0. The hosted Mastra platform (Studio, Server, Memory Gateway) is a paid cloud product with a free Starter tier.
How much does Mastra cost?
The framework is free. The hosted platform has a free Starter tier, a $250/month Teams tier, and custom Enterprise pricing as of June 2026.
Executive Summary
Mastra is an open-source TypeScript framework for building AI-powered applications and agents. Created by the founders of Gatsby and backed by Y Combinator (W25), it provides a batteries-included approach: agents, workflows, RAG, memory, MCP servers, and evals in one cohesive package. The framework hit v1.0 stable in January 2026,[1] and in April 2026 the company raised a $22M Series A led by Spark Capital ($35M total) while launching a hosted platform — Studio, Server, and Memory Gateway.[2] The standout feature remains Observational Memory, which achieves state-of-the-art benchmark scores without RAG or knowledge graphs.
| Attribute | Value |
|---|---|
| Company | Mastra (Gatsby founders: Sam Bhagwat, Abhi Aiyer, Shane Thomas) |
| Founded | 2024 |
| Funding | $35M total — YC W25, $13M seed (Oct 2025), $22M Series A led by Spark Capital (Apr 2026)[2] |
| Employees | ~10 (as of Feb 2026) |
| Headquarters | San Francisco, CA |
Product Overview
Mastra is an open-source TypeScript framework for building AI-powered applications and agents.[3] Created by the team behind Gatsby, Mastra is designed around the principle that "Python trains, TypeScript ships."[4]
The framework supports 90+ model providers through a unified interface — OpenAI, Anthropic, Google, DeepSeek, and more.[3] As of June 2026, the GitHub repo has ~25k stars,[4] and named users include Replit, Sanity, SoftBank, WorkOS, Brex, and Factorial.[2]
Key Capabilities
| Capability | Description |
|---|---|
| Agents | Autonomous agents that use LLMs and tools for open-ended tasks |
| Workflows | Graph-based orchestration with .then(), .branch(), .parallel() |
| Observational Memory | SOTA memory system (94.87% on LongMemEval) |
| RAG | Built-in retrieval from APIs, databases, and files |
| MCP | Model Context Protocol server authoring |
Product Surfaces / Editions
| Surface | Description | Availability |
|---|---|---|
| Framework | TypeScript npm package (v1.0 stable Jan 2026; @mastra/core 1.41.x as of June 2026) | GA |
| CLI | npm create mastra@latest | GA |
| Studio | Observability, collaboration, and evaluation tools (cloud or self-hosted) | GA (Apr 2026)[2] |
| Server | Hosted agent deployment | GA (Apr 2026) |
| Memory Gateway | AI gateway with managed agent memory | GA (Apr 2026) |
| Enterprise | Self-hosted enterprise features (RBAC, SSO, IAM) and support | Contact sales |
Technical Architecture
Language: TypeScript (Node.js)
Storage adapters: PostgreSQL, LibSQL, MongoDB
Key Technical Details
| Aspect | Detail |
|---|---|
| Deployment | Self-hosted (Node.js) or Mastra Server (hosted) |
| Model(s) | 90+ providers (OpenAI, Anthropic, Google, etc.)[3] |
| Integrations | Vercel AI SDK, CopilotKit, React, Next.js |
| Open Source | Yes (Apache 2.0) |
Installation:
npm create mastra@latest
Observational Memory (SOTA)
Mastra's standout feature is Observational Memory (OM), achieving state-of-the-art on LongMemEval:[5][6][7]
| System | Model | LongMemEval |
|---|---|---|
| Mastra OM | gpt-5-mini | 94.87% |
| Mastra OM | gemini-3-pro-preview | 93.27% |
| Supermemory | gpt-4o | 81.60% |
| Full context | gpt-4o | 60.20% |
OM uses background agents to compress conversations into dense observations, achieving 5-40× compression without RAG or graphs.[8] As of June 2026 the 94.87% result remains the highest full-benchmark score Mastra cites, though competing memory vendors (ByteRover, MemPalace) have since published their own self-reported LongMemEval results — treat all vendor benchmarks with the usual caution.
Strengths
- SOTA memory — Observational Memory leads published LongMemEval results
- v1.0 stable — Hit 1.0 in January 2026 with API stability guarantees[1]
- TypeScript-native — First-class TS experience, not a Python port
- Batteries included — One framework for agents, workflows, RAG, memory, evals
- Prompt cache friendly — Stable context windows enable cost savings
- Gatsby pedigree — Team has shipped widely-adopted open source before
- Well funded — $35M raised; $22M Series A led by Spark Capital (Apr 2026)[2]
- Human-in-the-loop — Built-in suspend/resume for approval workflows
- Production proof points — Deployments at Brex, Indeed, and Marsh McLennan (100k+ daily users)[2]
Cautions
- TypeScript only — Python developers have to switch stacks
- Young project — v1.0 only landed January 2026; production track record is growing but short vs. established frameworks
- OM sync limitation — Observation blocked conversation as of Feb 2026 (async mode promised)
- Gatsby ghost — Gatsby's decline after Next.js dominance raises questions about staying power
- Memory vendor lock-in — OM requires Mastra's storage adapters; not portable
- Platform metering — Hosted platform adds usage-based fees (observability events, CPU hours, egress, memory tokens) that can stack up[9]
Pricing & Licensing
| Tier | Price | Includes |
|---|---|---|
| Open Source | Free | Full framework (Apache 2.0), self-host anywhere |
| Platform Starter | Free | 100K observability events, 24 CPU hours, 15-day retention, 1GB storage, unlimited users |
| Platform Teams | $250/month | 1M events, 250 CPU hours, 6-month retention, 100GB storage, SSO, SOC 2 docs |
| Enterprise | Custom | RBAC, audit logs, SLAs, dedicated support; self-hosted option with flat annual fee |
Pricing as of June 2026.[9]
Licensing model: Open source (Apache 2.0) framework + usage-metered hosted platform + enterprise services
Hidden costs: Overage metering on the platform — $10/100K observability events (Starter), $0.35/CPU hour, $0.10/GB egress, $100/project for persistent 24/7 server uptime; Memory Gateway charges market token rates +5.5% and $20/GB retrieval storage[9]
Competitive Positioning
Direct Competitors
| Competitor | Differentiation |
|---|---|
| LangChain | LangChain is Python-first; Mastra is TypeScript-native |
| CrewAI | CrewAI focuses on agent teams; Mastra is full framework |
| Vercel AI SDK | Vercel is lower-level; Mastra uses it internally |
| Supermemory | Mastra OM outperforms Supermemory on benchmarks |
When to Choose Mastra Over Alternatives
- Choose Mastra when: You want TypeScript-native, batteries-included framework with SOTA memory
- Choose LangChain when: You're Python-first or need maximum modularity
- Choose CrewAI when: You specifically need multi-agent team orchestration
- Choose Vercel AI SDK when: You want lower-level primitives, not full framework
Ideal Customer Profile
Best fit:
- TypeScript developers building AI applications
- Teams wanting all-in-one framework without stitching libraries
- Projects needing production-ready memory that scales with conversation
- React/Next.js developers wanting familiar patterns
- Open source advocates wanting commercial backing
Poor fit:
- Python-first teams
- Organizations needing proven production track record
- Teams wanting to avoid any vendor lock-in
- Projects requiring immediate async memory operations
Viability Assessment
| Factor | Assessment |
|---|---|
| Financial Health | Strong — $35M raised, Series A led by Spark Capital (Apr 2026) |
| Market Position | Leading TypeScript agent framework — challenger to Python incumbents |
| Innovation Pace | Rapid — v1.0, SOTA memory, full hosted platform within 18 months |
| Community/Ecosystem | Growing — ~25k GitHub stars (June 2026), Gatsby team reputation |
| Long-term Outlook | Positive — funded, differentiated, with enterprise proof points |
Mastra now has $35M in funding, a stable v1.0, and named enterprise deployments (Indeed, Marsh McLennan, Brex).[2] The Gatsby team has proven open-source execution. The remaining risk is monetization — whether the metered platform converts the open-source base — and avoiding Gatsby's fate of being eclipsed by competitors.
Bottom Line
Mastra is the most complete TypeScript AI framework available. The Observational Memory system is genuinely novel — achieving SOTA without RAG or graphs is impressive, and the prompt-caching benefits have real cost implications at scale.
The Gatsby team knows how to build developer tools that get adopted. Whether Mastra becomes the "Next.js of AI frameworks" depends on TypeScript developer adoption.
Recommended for: TypeScript developers wanting a batteries-included AI framework with SOTA memory and React/Next.js integration.
Not recommended for: Python-first teams, organizations needing proven production track record, or projects avoiding any framework lock-in.
Outlook: With v1.0 shipped, $35M raised, and enterprise deployments at Indeed and Marsh McLennan, Mastra is the front-runner to become the default TypeScript AI framework. Watch platform monetization and whether Observational Memory holds its benchmark lead as memory vendors crowd in.
Research by Ry Walker Research • methodology