LagrangeCloud
Get started

Ship your agent in one sentence. Better with every run.

Deploy the agent you built in Codex or Claude Code, or describe a new one and AI builds it. Say what to adjust any time, and Lagrange keeps making it cheaper and faster.

Install the plugin Open the console
Codex · ~/acme/support-bot

Deploy this agent to LagrangeCloud on Lagrange

  1. Read AGENTS.md, 2 skills and a script
  2. Added MCP servers linear and stripe
  3. Stored 2 keys as secrets, without printing them
  4. Deployed support-bot v1 on Lagrange

Live https://d3iwapvcygehqg.cloudfront.net/v1/invoke/ag_5d02c7a3b81e

Acme supportlive

Where is order 42?

It shipped yesterday and arrives Friday.

Lagrange found an optimizationLook up each order once Cost per run$0.031$0.013 Median time4.8s2.0s Promoted. support-bot v2 is live.

Lagrange RSI Harness. Every run makes the next one better.

Most agents stay as good as the day they shipped. Lagrange reads the trace of every run, finds a cheaper or faster way to do the same work, tests it on live traffic and promotes it once you approve. Then it starts again from the new version.

support-bot Watching v1 Pending review Canary 10% Promoted
  1. Observe
  2. Propose
  3. Review
  4. Canary
  5. Promote

Trace of a typical run

Most runs look up the same order 3 times.

Look up each order once

instructions:
+ Call lookup_order once and reuse the result.

v1 and v2 on live traffic

v1v2

Cost per run $0.031 $0.013
Median time 4.8s 2.0s
Success rate 96% 97%
v2 is serving 10% of runsv2 is serving all runs. Looking for v3.

Everything it uses comes with it. Keys go to secrets, not the chat.

~/acme/support-bot 
AGENTS.md
skills/orders/SKILL.md
skills/refunds/SKILL.md
scripts/lookup_order.py
.codex/config.toml  linear, stripe
.env                LINEAR_API_KEY, STRIPE_SECRET_KEY
support-botv1 live
Instructions
AGENTS.md
Skills
orders, refunds
Files
lookup_order.py
MCP servers
linear, stripe
Secrets
LINEAR_API_KEY, STRIPE_SECRET_KEY
Harness
Lagrange or Codex, Claude Code

One deploy. Three ways to use it.

From your app

client = OpenAI(
    base_url=endpoint + "/openai",
    api_key=caller_key,
)
client.chat.completions.create(
    model="support-bot",
    messages=history,
)

On your site

acme.com
Acme support

Can I return these boots?

Yes, within 30 days. I've started the return.

On a schedule

0 8 * * *Triage new tickets
  • Today 08:0041s
  • Yesterday 08:0038s
  • Tuesday 08:0044s

Every step, traced.

run_3b81e6d02c7a succeeded
2m14s · $0.27 · 66.0k tokens
  1. run · lagrange2m14s
  2. plan0.3s
  3. model1.4s
  4. tool · shell18.2s
  5. subagent · research34.6s
  6. model12.1s
  7. tool · grep0.8s
  8. model19.9s
  9. context compacted0.4s
  10. model9.4s
  11. tool · edit1.1s
  12. tool · shell58.9s
  13. verify · pass7.4s
model tool subagent harness verify Example trace

Pay for what runs.

Trial credit

$5

Runs, instances and fleets

from$0.05per sandbox hour

Plus model tokens at list price.

Try it on your agent.

$ npx -y @openai/codex plugin marketplace add https://github.com/TensorBlock/lagrange-plugins
$ npx -y @openai/codex plugin add lagrange@tensorblock
Then, in your agent's folder: "Deploy this agent to LagrangeCloud"
Get started Sign in