Beam
Also known as: Beam Cloud, beam.cloud, beta9
Serverless cloud for AI workloads with GPU inference, sandboxes, task queues and an agent framework, running on Beam's cloud, your own AWS or GCP account, or self hosted from the open source Beta9 engine.
Beam is a serverless cloud for AI and ML workloads, with GPU inference, task queues, sandboxes and scheduled jobs defined in Python or TypeScript and scaled from zero. Beta9, the engine that powers Beam, is open source under AGPL 3.0 and can be self hosted on a local machine or Amazon EKS. The SDK and CLI switch between the two by changing an import.
Sandboxes run any code with GPUs, networking and storage, boot in one to three seconds, and snapshot their filesystem and memory to fork or restore. Scheduled jobs, endpoints that boot from zero, task queues with retries and callbacks, and signals between apps start and chain work. Beam's Bot agent framework models an agent as a Petri net of states and transitions that run concurrently in isolated containers. Volumes, durable disks, distributed queues and maps, and S3 mounts hold data.
Beam runs on AWS and Google Cloud, in the customer's own AWS or GCP account through bring your own cloud, or on compute pools of the customer's own machines. It has a SOC 2 Type II audit by Advantage Partners, HIPAA BAAs for Enterprise and a GDPR DPA. Single sign on is not yet available, and all workspace members share the same permissions.
Pricing is usage based on top of a plan. Developer has no monthly fee, Team costs 89 dollars a month and Growth is priced by contact. Sandboxes start from about 32 cents an hour and H100s from 1.83 dollars an hour, storage is free to a terabyte and there are no egress charges.
Vendor details
Canonical URL
https://www.beam.cloud
Category
Agent infrastructure
Subcategory
Compute and sandboxes
Funding status
Independent. The Beta9 engine that powers Beam is open source under AGPL 3.0 (github.com/beam-cloud/beta9, actively released), with a managed cloud and bring your own cloud on AWS or GCP.
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
Python and TypeScript SDKs, a CLI, a REST API with Swagger specs, web endpoints, task queues with callbacks, scheduled jobs, signals between apps, volumes, durable disks and S3 mounts, custom registries, and the Bot agent framework.
In practice
Your long horizon coding agent must keep state across many actions. Beam gives sandboxes that run any code with GPUs and snapshot their filesystem and memory, so the agent can fork or restore its state and carry on.
Sensitive code cannot leave your cloud. You self host the open source Beta9 engine on a local machine or Amazon EKS, or let Beam run the control plane and compute inside your own AWS or GCP account with bring your own cloud, so the compute bills to your existing cloud credits.
Your agent spins up sandboxes constantly and cold start billing hurts. Beam does not charge for server spin up or container image load time, only for loading your code, and bills by the millisecond, so the high churn pattern stays cheap.
Sources & related URLs
Agentic Index coverage score
7.5 / 14 capabilities · 54%
| Integrations & Tool Calling | Partial |
|---|---|
|
Code on Beam reaches other systems through what the customer writes. Secrets hold credentials, S3 buckets mount into apps and custom registries supply images. Task callbacks POST results to the customer's server, and sandbox ports expose services. Beam has no catalog of connectors, OAuth actions or tool gateway of its own, and the tools are the customer's code. Sourcedocs.beam.cloud/llms.txtread 2026-09-22 |
|
| Workflow Orchestration | Full |
|
An agent framework ships in the SDK (Bot, BotLocation, BotContext) and is built on a Petri net model. Locations hold typed state, and transitions consume inputs from locations and write outputs to others. Transitions can be driven by a model or by plain code, run concurrently in isolated containers and keep state synchronized between them. Task queues add retries and timeouts, and map fans work out. Deterministic steps mix with agent steps, and branching and retries are built in. Beam says it is launching the framework and does not say how mature it is beyond the example it gives. Sourcedocs.beam.cloud/v2/agents/introductionread 2026-09-22 |
|
| Knowledge Grounding & RAG | Not documented |
|
Any retrieval system on Beam is one the customer builds. Beam maintains no retrieval structure of its own over the customer's content for an agent to query. Volumes, durable disks, Maps and mounted S3 buckets store data but return no search results, and nothing Beam provides indexes, refreshes or permissions the customer's sources. Sourcedocs.beam.cloud/llms.txtread 2026-09-22 |
|
| Human Oversight & Guardrails | Not documented |
|
For work running on Beam, including agents built with its framework, Beam names no approval step, consent checkpoint or escalation to a person. The platform isolates workloads with containers per customer, network separation and scoped secrets. Sourcedocs.beam.cloud/v2/security/securityread 2026-09-22 |
|
| Security, Identity & Governance | Partial |
|
A SOC 2 Type II audit by Advantage Partners covers Beam, with the report under NDA, and Beam offers HIPAA BAAs for Enterprise customers and a GDPR DPA. Data is encrypted in transit and at rest, and containers and networks are isolated per customer. The customer scopes, rotates and deletes its encrypted secrets, and workspace API tokens can be created, rotated and revoked. All members of a workspace currently have the same permissions, and single sign on is not yet available. Sourcedocs.beam.cloud/v2/security/securityread 2026-09-22 |
|
| Observability & Auditability | Partial |
|
Application logs are kept 30 days on Developer and Team plans and one year on Growth, and log export is included. Task status can be queried by ID, with callbacks on completion. Prompts, tool calls, retrieved knowledge and outputs can be inspected step by step only as far as the customer's code logs them. Beam names no audit log of workspace actions separate from the runtime logs. Sourcebeam.cloud/pricingread 2026-09-22 |
|
| Memory & State Persistence | Partial |
|
State carries across runs. Sandbox snapshots capture filesystem and memory as an immutable artifact to fork from or restore, and distributed volumes and durable disks on each node persist data. Distributed Queues and Maps share it across tasks, and the agent framework synchronizes state between transitions, so an agent's code can read that state and carry on. Snapshot state is the machine's own state, which a customer can restore or delete but not review or edit. Beam names no memory layer for agents. Sourcedocs.beam.cloud/v2/sandbox/snapshotsread 2026-09-22 |
|
| Deployment & Data Residency | Full |
|
There are three ways to run Beam. The managed cloud runs on AWS and Google Cloud. With bring your own cloud, Beam runs the control plane and compute in the customer's AWS or GCP account for a flat management fee, and the compute bills to the customer's cloud. The pricing plans add Azure as a bring your own cloud option. Beta9, the engine that powers Beam, is open source under AGPL 3.0 and can be self hosted on a local machine or Amazon EKS. Compute pools can also run Beam workloads on the customer's own machines. The beta9 repository is actively maintained. Beam names no region selection for the managed cloud. Sourcedocs.beam.cloud/v2/self-hosting/overviewread 2026-09-22 |
|
| Prebuilt Agents, Templates & Packs | Partial |
|
Ready made starting points ship as end to end examples for model serving (Llama, DeepSeek, Qwen and Hugging Face models, plus an OpenAI compatible vLLM server), for media (ComfyUI, Whisper and text to speech), and for fine tuning and web scraping, plus a research assistant agent built on the agent framework. Developers build from these examples, and Beam offers no browsable catalog of agents a customer can adopt as they are. Sourcedocs.beam.cloud/v2/examples/overviewread 2026-09-22 |
|
| Triggers & Channel Coverage | Full |
|
The schedule decorator runs a deployed job on a schedule, and web endpoints and task queues boot containers from zero when requests or tasks arrive. One app can send events to another through signals, for example to reload a model when retraining finishes. Task callbacks also notify the customer's server when a task finishes. Sourcedocs.beam.cloud/v2/function/scheduled-jobread 2026-09-22 |
|
| Model Flexibility & Routing | Full |
|
Customers choose the model, and Beam runs any model on the GPU the customer picks. Examples serve Hugging Face models such as Llama 3.1, DeepSeek R1 and Qwen through vLLM, including an OpenAI compatible vLLM server, or through SGLang, and others fine tune Gemma and Llama. Customers can bring their own deployments and providers. Sourcedocs.beam.cloud/llms.txtread 2026-09-22 |
|
| APIs, SDKs & MCP Extensibility | Full |
|
Outside callers reach Beam through Python and TypeScript SDKs, a CLI, and a REST API with published Swagger specifications for the gateway and pods, covering tasks, pods and sandboxes. Web endpoints expose the customer's code as HTTP APIs, the open source Beta9 engine can be extended, and a docs MCP server serves Beam's documentation to coding tools. Sourcedocs.beam.cloud/llms.txtread 2026-09-22 |
|
| Testing, Debugging & Optimization | Not documented |
|
Code, experiments and fine tuning jobs run on Beam with retries and logs. Beam names no fixtures, scoring or quality gates for a customer's agent workflows, and no way to compare agent runs. Sourcedocs.beam.cloud/llms.txtread 2026-09-22 |
|
| Browser & Computer Use | Not documented |
|
Sandboxes run any code with GPUs, networking and storage and can expose ports for web services. Beam names no browser or desktop for an agent to operate and no API for screenshots, clicks or typing. Sourcedocs.beam.cloud/v2/sandbox/overviewread 2026-09-22 |
|
The Agentic Index coverage score grades every vendor Full, Partial or Not documented against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Pricing
Developer $0 plus usage · Team $89/mo plus usage · sandboxes from $0.319/hr
per second and per millisecond compute (GPU, CPU, RAM)
Included quota
Developer: 1 seat, 30-day log retention. Team at $89 a month: 3 seats ($25 per additional seat), 30-day log retention. Growth: unlimited seats, one-year log retention. Storage up to 1 TB and egress are included on every plan; compute bills per second or millisecond.
What is public
Beam publishes its plans, per second serverless, sandbox and on-demand rates, GPU rates, storage and egress terms, and the BYOC management fee.
Billing mechanics
Pay per millisecond of active runtime with scale to zero, no billing during cold starts or image pulls, and storage included free. Managed cloud price is a ceiling because the open source core supports self hosting and bring your own compute on your own AWS, GCP, Azure, or Hetzner credits.
Cost watchouts
Serverless CPU and RAM cost more than the on-demand rates, and sandboxes bill while running; BYOC adds a management fee on top of your own cloud bill; storage beyond 1 TB bills per GB-month.
Variable cost rationale
No subscription floor for production; cost is per millisecond compute that scales with GPU, CPU, and RAM runtime, though self hosting and bring your own compute let a team cap the managed price.
Overage / add-ons
Compute bills per second or millisecond above any included usage across GPU, CPU and RAM; storage beyond 1 TB at $0.021 per GB-month; volume discounts above $10K a month.
Sales call required
No, self serve available
Free / trial
Developer plan at $0 a month plus usage; the open source Beta9 engine is free to self host
Lowest paid plan
Team $89 a month plus usage
Commercial notes
Independent. Open source engine beta9 under AGPL. Differentiates on stateful, GPU capable sandboxes with bring your own compute for production agents.
Key ambiguities
Rates differ between serverless, sandbox, on-demand and BYOC compute, and GPU rates vary by card; confirm the rate for the workload shape you run.
Missing data
Growth pricing is by contact; any monthly credit on the Developer plan was not stated on the pages read.
Related vendors
- AgentOps — Agent observability and debugging platform: open source SDKs trace…
- AgentPhone — AgentPhone gives the AI agents a team builds their own US or…
- Agno — Python agent framework and AgentOS runtime (formerly Phidata) for…
- AIsa — Resource and payment gateway for AI agents, with one key to 110+…
- Albato — Albato is a no code integration platform that sells an embedded,…
- AlphaBitCore — AI control plane for regulated financial firms: one gateway enforces…
Alternatives to Beam
The closest documented capability profiles to Beam among agent infrastructure platforms tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Modal8.0 / 14Adds documented Browser & Computer UseBeam vs Modal →
- E2B8.0 / 14Adds documented Browser & Computer UseBeam vs E2B →
- GPT Researcher7.0 / 14Adds documented Knowledge Grounding & RAG and Browser & Computer Use
- Inngest10.0 / 14Adds documented Human Oversight & Guardrails and Testing, Debugging & Optimization
- Vocode6.0 / 14Fuller documented coverage on Integrations & Tool Calling
- Bernstein10.5 / 14Adds documented Human Oversight & Guardrails and Testing, Debugging & Optimization
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded