Blaxel
Blaxel provides cloud infrastructure — microVM-based sandboxes, agent runtime, batch jobs, MCP hosting, and a model gateway — purpose-built for deploying and scaling autonomous AI agents, primarily serving AI agent developers and startups via a usage-based, developer-first platform.
- Company typePrivate
- Founded2024
- HeadquartersSan Francisco, United States
- Headcount1–10
- GTM typeB2B
- OfferingSoftware
What Blaxel does
Blaxel, Inc. is a San Francisco-based cloud infrastructure company building compute, storage, and networking primitives purpose-built for autonomous AI agents. Founded in 2024 as Beamlit and rebranded to Blaxel in February 2025, the company operates from 945 Market St and is backed by Y Combinator (W25) and First Round Capital. Its core platform consists of microVM-based Sandboxes that boot in milliseconds and resume in ~25ms with persistent memory state, a session-first Agent Runtime for long-running stateful agents, Batch Jobs that fan out across thousands of sandboxes, MCP server hosting on a private network, Agent Drive for distributed multi-agent filesystem collaboration, and a Model Gateway providing unified access to external AI model providers.
The technology is delivered through Python, TypeScript, and Go SDKs, a CLI, and a REST API, and is offered as managed cloud, self-hosted, or VPC deployments across four regions (Oregon, Virginia, London, Frankfurt). The go-to-market is product-led with $200 in free credits for self-serve onboarding, developer-first documentation, open-source presence on GitHub, and a Discord community; enterprise prospects engage through a direct engineering sales motion. Revenue is generated via usage-based pricing on compute runtime, storage (snapshots, volumes, images), and custom domain active time, with $0 compute cost while sandboxes are idle. Named customers include Webflow and Strapi as enterprise references alongside a long tail of early-stage AI agent startups (Shortwave, Polsia, Sapiom, Tasklet, Vybe, Mendral, Casco, Ploy, Cubic, Bloom, Cartage, HireThomas), and the company has secured SOC 2 Type II, HIPAA, and ISO 27001 certifications to support regulated workloads.
Blaxel firmographics
Firmographics- Name
- Blaxel
- Legal name
- Blaxel, Inc.
- Website
- https://blaxel.ai
- Company type
- Private
- Founded year
- 2024
- Operating status
- Operating
- Headcount range
- 1–10 employees
- Short description
- Blaxel provides cloud infrastructure — microVM-based sandboxes, agent runtime, batch jobs, MCP hosting, and a model gateway — purpose-built for deploying and scaling autonomous AI agents, primarily serving AI agent developers and startups via a usage-based, developer-first platform.
- Ownership category
- akta.pro rank
Blaxel industry classification
Industry- Product category
- AI Agent Cloud Infrastructure
- NAICS
- Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (5182), Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821), Software Publishers (513210)
- SIC
- Services-Computer Programming, Data Processing, Etc. (7370), Services-Prepackaged Software (7372)
- akta.pro primary industry
- AI Compute Virtualization & Scheduling (GPU virtualization, cluster schedulers) (HDAAAAAG)
- akta.pro secondary industries
- AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs) (HDAEANAJ), Model Serving, Inference & Deployment Platforms (APIs, Edge/On-Prem) (HDAEANAC), End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA), iBPMS (Intelligent BPM Suites) & Process Automation Suites (HDAEAHAG)
Keywords
Where Blaxel is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Offices1 record
Markets served
Blaxel business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Technology or R&D, Infrastructure, Personnel, Marketing or Sales, Operations
Revenue model
- Compute Runtime: Usage-based billing for compute runtime measured in GB-seconds (gbs). Sandboxes are billed for compute time with per-second rates offering up to 70% lower costs than traditional cloud providers in practice.
- Storage: Charging for storage costs including Sandbox Snapshots, Sandbox Volumes, and stored deployment Images. Storage is measured in GB-seconds.
- Custom Domain Active Time: Billing for custom domain usage measured in active hours when domains are deployed and routing traffic.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Usage-based | Pay-as-you-go | Free credits tier for new users to get started |
| Usage-based | Pay-as-you-go | Usage-based compute and storage |
Go-to-market motion1 record
Distribution channels2 records
Marketing channels6 records
Blaxel product offering
Product offeringCore offering
Blaxel provides microVM-based cloud infrastructure specifically designed for deploying and scaling autonomous AI agents. The platform delivers compute (Sandboxes, Agent Runtime, Batch Jobs, MCP Servers), storage (Agent Drive, Volumes, Sandbox Local Filesystem), and networking (Outbound Control, Static IPs, Proxy Routing, Model Gateway) as built-in primitives, enabling developers to ship AI agents without reinventing infrastructure. Customers access the platform via self-serve (app.blaxel.ai), SDKs (Python, TypeScript, Go), CLI, and REST API, with deployment options including managed cloud, self-hosting, and VPC.
Product overview
Blaxel is an AI agent infrastructure platform providing compute, storage, and networking primitives designed specifically for autonomous AI agents. The platform consists of three core product areas: Compute (Sandboxes, Agent Runtime, Batch Jobs, MCP Servers), Storage (Agent Drive, Volumes, Sandbox Local Filesystem), and Networking (Outbound Control, Static IPs, Proxy Routing, Model Gateway). Sandboxes are microVM-based environments that boot in milliseconds with persistent memory state, enabling AI agents to reason, code, and act in isolated environments. Agent Drive provides distributed filesystem for multi-agent collaboration. The Model Gateway unifies access to multiple LLM providers (OpenAI, Anthropic, DeepSeek, HuggingFace, Azure AI Foundry, Cohere, xAI, Mistral AI). Developer tools include Python, TypeScript, and Go SDKs plus a CLI. Deployment options include managed cloud, self-hosting, and VPC. The platform is rebranded from Beamlit (formerly Beamlit) and is Y Combinator-backed. Pricing is usage-based with $0 compute cost while idle.
Differentiator
Problem solved
Functional benefit
Products and services
- Sandboxes MicroVM-based isolated compute environments for AI agents that boot in milliseconds and resume in ~25ms with memory state intact. One microVM per agent, app, or job with persistent environments that stay on standby indefinitely when idle.
- MCP Servers Model Context Protocol server hosting for running MCP servers as first-class private workloads on the same private network as agents. Supports pre-built integrations (Tavily, Supabase, HubSpot, Smartlead, AgentMail, Context7, Linkup, Airweave, Nia, Brave Search, Google Maps, Slack, Linear, AWS S3, AWS SES, PostgreSQL, Cloudflare, Dall-E) and custom MCP server deployment.
- Agent Drive Distributed filesystem for real-time multi-agent collaboration. Enables persistent shared memory across sandboxes and agents with read-write or read-only mounting options and shared session context and state across agents.
- Batch Jobs Distributed task execution system that spawns thousands of tasks across individual sandboxes in seconds for parallel workload execution. Supports cron-based triggers, job revisions, ephemeral volumes, and use as GitHub Runners.
- Volumes Durable storage mounts that retain data for years with redundant backend. Can be resized dynamically and attached to agents or sandboxes for persistent file storage.
- Static IPs Dedicated outbound IP addresses for sandboxes enabling IP whitelisting at external services. Also available as dedicated egress gateways in private preview.
- Model Gateway Unified endpoint providing co-located access to every model provider. Centralizes model access, credentials, and observability behind a production-ready gateway with automatic routing optimization across OpenAI, Anthropic, DeepSeek, HuggingFace, Azure AI Foundry, Cohere, xAI, and Mistral AI.
- Blaxel SDK Software development kits for Python, TypeScript, and Go enabling programmatic access to all Blaxel resources including sandboxes, agents, jobs, and model APIs. Provides helpers for connecting to model providers and tool servers hosted on Blaxel.
- Blaxel CLI Command-line interface for managing Blaxel resources including sandbox creation, job deployment, agent deployment, and MCP server management. Supports 'bl new' for quick project initialization and 'bl deploy' for deployment, plus a CLI skill usable by coding agents.
- Blaxel REST API REST API for managing all Blaxel resources including agents, applications, sandboxes, jobs, volumes, and MCP servers. Supports API key and OAuth 2.0 authentication with header-based versioning using Stripe-style date format and cursor-based pagination.
- Blaxel Console Web-based management console for Blaxel with workspace management, resource monitoring, billing explorer, logs viewer, and integrated changelog.
Quantifiable outcome
- 25ms resume time for sandboxes
- +3 more outcomes
Companies that use Blaxel
Customer profileNamed customers14 records
Segments2 records
Ideal customer profiles2 records
Blaxel technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration41 records
AI capability9 records
Feature10 records
Blaxel partnerships and signals
Strategic signalPartnerships
Eight partnerships are on record, tiered core.
- Anthropic (Claude)coreIntegration with Claude Managed Agents allowing self-hosting of the sandbox execution layer while keeping the agent loop within Anthropic's platform. Every tool call runs inside a Blaxel sandbox.
- LangChain/LangGraphcoreNative SDK support with automatic tool conversion for LangChain and LangGraph frameworks.
- CrewAIcoreNative SDK support with automatic tool conversion for CrewAI multi-agent framework.
- OpenAIcoreIntegration with OpenAI Agents SDK including streaming, file transfer, and Next.js code generation capabilities.
- GooglecoreSupport for Google ADK (Agent Development Kit) framework for building and deploying agents.
- VercelcoreNative support for Vercel AI SDK integration.
- PydanticAIcoreNative SDK support with automatic tool conversion for PydanticAI agent framework.
- LlamaIndexcoreNative SDK support with automatic tool conversion for LlamaIndex framework.
Scale indicators7 records
Recent moves7 records
Expansion highlights7 records
Blaxel competitors and assessment
Company assessmentDirect peers
- Modal Labs: Modal provides serverless compute infrastructure for AI/ML workloads with on-demand containers, GPU support, and Python-first SDKs. It is the most directly comparable cloud infrastructure platform for running AI code and models, competing head-to-head with Blaxel's Sandboxes on performance and developer experience.
- E2B: E2B offers open-source secure sandboxed cloud environments purpose-built for AI agents and code execution. It is the closest direct competitor to Blaxel's Sandboxes product — Blaxel even offers explicit "migration skills" to help customers move from E2B to Blaxel, signaling head-to-head positioning.
- Replicate: Replicate is a cloud platform for running and scaling AI models via API, with a focus on serving open-source and custom ML models in production. It overlaps with Blaxel's Model Gateway and Sandbox products for AI inference and execution workloads.
- Fly.io: Fly.io runs full-stack applications and processes close to users via lightweight VMs (Firecracker-based) across global regions. It competes with Blaxel on distributed application hosting, microVM infrastructure, and developer-friendly global deployment.
- Railway: Railway provides a developer-friendly PaaS for deploying applications with usage-based pricing, built-in databases, and managed infrastructure. It overlaps with Blaxel's deployment, compute, and storage primitives for backend workloads serving AI applications.
- Anyscale: Anyscale, built on the open-source Ray framework, provides a managed platform for distributed AI/ML compute and LLM serving. It competes with Blaxel's Batch Jobs and Model Gateway capabilities for production AI workloads at scale.
- Render: Render is a unified cloud platform for hosting web services, APIs, databases, and cron jobs with a developer-first experience and usage-based pricing. It overlaps with Blaxel's broader PaaS offering for AI-driven applications and agents.
Broad incumbents
- Cloudflare Workers: Cloudflare Workers offers a globally distributed serverless compute platform with V8 isolates, durable objects, and Workers AI. Blaxel already integrates with Cloudflare as a sandbox provider, and Cloudflare competes at the edge-compute layer for agent execution.
- AWS (Bedrock AgentCore / Lambda): AWS provides broad cloud infrastructure including Bedrock AgentCore and Lambda for AI agent and serverless workloads. As a hyperscaler with bundled pricing, existing enterprise relationships, and multi-billion-dollar capex, AWS is the dominant incumbent Blaxel must displace or partner with.
Emerging players
- Vercel: Vercel offers a frontend cloud platform with native AI SDK integration, sandbox provider capabilities, and edge functions. It is a partial competitor for AI-driven web/agent application deployment and is also a Blaxel integration partner.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat5 records
Key risks2 records
Key highlights7 records
Customer concentration
Blaxel social profiles
Digital presenceBlaxel compliance and trust
Trust signalCompliance3 records
Blaxel financial estimates
Financial estimateRevenue estimate
Valuation estimate
Blaxel leadership team
Management profileNumber of profiles
Profiles6 records
Blaxel funding detail
Funding detailFunding overview
Funding rounds1 record
Investors7 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Blaxel M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about Blaxel
What does Blaxel do?
Blaxel provides microVM-based cloud infrastructure specifically designed for deploying and scaling autonomous AI agents. The platform delivers compute (Sandboxes, Agent Runtime, Batch Jobs, MCP Servers), storage (Agent Drive, Volumes, Sandbox Local Filesystem), and networking (Outbound Control, Static IPs, Proxy Routing, Model Gateway) as built-in primitives, enabling developers to ship AI agents without reinventing infrastructure. Customers access the platform via self-serve (app.blaxel.ai), SDKs (Python, TypeScript, Go), CLI, and REST API, with deployment options including managed cloud, self-hosting, and VPC.
Is Blaxel a public or private company?
Blaxel is a private company. It is classified as venture growth investor backed and is currently operating.
When was Blaxel founded?
Blaxel was founded in 2024. It employs 1 to 10 people.
Where is Blaxel based?
Blaxel is headquartered in San Francisco, United States, in the North America region.
How does Blaxel make money?
Three revenue lines are on record. Compute Runtime is the primary driver. The others are storage and custom Domain Active Time.
Who are Blaxel's main competitors?
Direct peers on record are Modal Labs, E2B, Replicate, Fly.io, Railway, Anyscale and Render. Broad incumbents are Cloudflare Workers and AWS (Bedrock AgentCore / Lambda). Vercel is listed as an emerging player.
Does Blaxel have an API?
Yes. Blaxel provides a REST API for managing all resources within workspaces, including agents, applications, sandboxes, jobs, volumes, MCP servers, and more. Authentication supports API keys (long-lived tokens) and OAuth 2.0 (short-lived 2-hour JWT tokens). The API uses header-based versioning with a Stripe-style date format (e.g., Blaxel-Version: 2026-04-16). Pagination is supported via cursor-based pagination for Management API list operations. Available endpoints include Management API, Inference API, and Billing API. MCP servers are available as first-class workloads on the same private network, and Blaxel provides Model Context Protocol (MCP) server hosting capabilities. Developer documentation is at docs.blaxel.ai/api-reference/introduction.
What industry is Blaxel in?
Blaxel's product category is AI Agent Cloud Infrastructure. Its primary akta.pro industry code is HDAAAAAG, AI Compute Virtualization & Scheduling (GPU virtualization, cluster schedulers), with a secondary code of HDAEANAJ, AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs). Its NAICS code is 5182 and its SIC code is 7370.