New Agent Runtime v2 is generally available

Ship AI agents that run in production.

Fermion is the runtime for agentic AI. Deploy tools, models, and long-running workflows behind one fast API — with sub-second cold starts, durable state, and observability built in.

No credit card required · 5,000 free agent runs every month
Scroll · build, run, observe
Build. Run. Observe.
agent.ts
TypeScript
1import { Fermion } from "@fermion/sdk";
2
3const agent = new Fermion.Agent({
4  model: "orion-4",
5  tools: [search, postgres, browser],
6  memory: "durable",
7});
8
9const run = await agent.stream({
10  input: "Draft the Q3 board update",
11});

Define once. Declare your agent — model, tools, and durable memory — in a few typed lines.

run · orion-4
LIVE
$ agent.stream({ input: "Draft the Q3 board update" })
plannerreasoning · plan readyok
retrieval12 docs · 302msok
tool·postgresSELECT metrics · 96msok
synthesisstreaming 412 tokens·
p95 latency -18%
412ms

Run anywhere. Stream long-running workflows behind one fast API, with sub-second cold starts.

Live traces
last 60 min
10:04p50 · 210msp95 · 412ms
planner180ms
retrieval302ms
synthesis418ms

Observe everything. Trace every tool call, token, and dollar in real time — then replay any run.

Trusted by engineering teams shipping AI to millions

Platform

Everything you need to run agents in production.

Stop stitching together queues, sandboxes, and dashboards. Fermion is the full runtime — from the first prototype to millions of runs a day.

Sub-second cold starts

Snapshot-restored sandboxes resume in 380ms, so idle agents still feel instant. Never pay for warm pools you don't need.

One API, every model

Route across frontier and open models with automatic failover, streaming, and per-token cost caps — no vendor rewrites.

Durable execution

Long-running workflows survive restarts, retries, and deploys. State is checkpointed, so nothing is ever lost mid-run.

Observability built in

Trace every tool call, token, and dollar in real time. Replay any run step by step and share a permalink with your team.

Type-safe SDKs

First-class TypeScript and Python with generated types for every tool. Go from a blank file to a deployed agent in an afternoon.

Secure by default

SOC 2 Type II, SSO, and private VPC networking. Your prompts and data are isolated per tenant and never train a model.

Observability

See exactly what your agents are doing.

Every run is a distributed trace. Drill into spans, inspect the inputs and outputs of each tool call, and catch regressions before your users ever notice them.

Live traces
last 60 min
10:04p50 · 210ms10:34p95 · 412ms11:04
planner180ms
retrieval302ms
tool·postgres96ms
synthesis418ms
380ms
Median cold start
99.98%
Rolling 90-day uptime
12B+
Agent runs per month
40+
Global edge regions
Pricing

Simple, usage-based pricing.

Start free and scale when you do. No seat taxes, no annual lock-in, no surprises on your bill.

Hobby
$0 /mo
For experiments, prototypes, and side projects.
Start for free
  • 5,000 agent runs / month
  • 2 concurrent sandboxes
  • 7-day trace retention
  • Community Slack support
Enterprise
Custom
For scale, compliance, and dedicated support.
Talk to sales
  • Volume pricing & committed use
  • Private networking (VPC peering)
  • SSO / SAML & 99.99% SLA
  • Dedicated solutions engineer

Deploy your first agent in minutes.

Free to start, sub-second cold starts, and full observability from your very first run.