Weights & Biases for AI agents

Weights & Biases tracks machine learning experiments, runs, and artifacts, used by ML teams comparing model training results across projects. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.

Connect Weights & Biases See the 10 actions ↓

Developer toolsAPI keyMCP + RESTwandb.ai

AUDIT LOG · WEIGHTS & BIASESPOLICY: acme-support

09:41:07 · claude · u_8f2read

weights_and_biases.list_runs

List Runs✓ allowed · 212ms

09:41:08 · claude · u_8f2read

weights_and_biases.get_team

Get Team✓ allowed · 164ms

EVERY WEIGHTS & BIASES CALL, ON THE RECORD

01 · USE CASES

What agents do in Weights & Biases.

List runs in a project

Pull recent training runs and their metrics as a read-only comparison.

Get a project's reports

Read saved reports summarizing experiment results and comparisons for a project.

List artifact types

Check what artifact types, like models or datasets, exist in a project.

02 · ACTIONS

10 Weights & Biases actions, graded by risk.

Every Weights & Biases action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.

read

Look things up. Allowed by default.

weights_and_biases.get_team
Get Team
weights_and_biases.list_runs
List Runs
weights_and_biases.list_sweeps
List Sweeps
weights_and_biases.list_reports
List Reports
weights_and_biases.list_projects
List Projects
weights_and_biases.get_organization
Get Organization
weights_and_biases.list_automations
List Automations
weights_and_biases.list_integrations
List Integrations
weights_and_biases.get_current_viewer
Get Current Viewer
weights_and_biases.list_artifact_types
List Artifact Types

write

Create and change things. Allow, or ask the user first.

No write actions.

destructive

Delete, cancel or archive. Ask first, or deny outright.

No destructive actions.

03 · HOW IT WORKS

Weights & Biases in three steps.

01Your users connect Weights & BiasesThey add their Weights & Biases api key on Arc0 Connect, under your brand. It goes straight into the vault.
02You set the rulesReads run, and writes like “updates” can wait for the user to approve.
03Any agent can actYour agent calls Weights & Biases through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.

POLICY.TS

await arc0.policies.set('weights_and_biases', {
  read: 'allow',
  write: 'ask',        
  destructive: 'deny',  
})

# Claude Code: the same connection, one URL
$ claude mcp add --transport http arc0 \
    https://mcp.arc0.ai/u/u_8f2

04 · AUTH AND DATA

How Weights & Biases connects.

Users add their Weights & Biases api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

The same Weights & Biases connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →

AUTH: API key
CREDENTIALS: Per-tenant encrypted vault
MODEL SEES: Results only, never credentials
AUDIT LOG: Every call, on every plan

05 · WORKS WITH

Use Weights & Biases from any agent.

Claude

ChatGPT

Cursor

Codex

VS Code

OpenAI Agents SDK

Claude Agent SDK

Vercel AI SDK

Mastra

LangGraph

06 · RELATED

More developer tools apps.

Addressfinder17 actions

Addresszen2 actions

Alchemy36 actions

ALL DEVELOPER TOOLS APPS →

07 · FAQ

Weights & Biases and Arc0, answered.

Q01

Can I use Weights & Biases with Claude, ChatGPT or Cursor?

Yes. Connect Weights & Biases to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the Weights & Biases actions you allow.

Q02

How do users connect Weights & Biases?

Users add their Weights & Biases api key on Arc0 Connect. It is encrypted in the vault, never shown to the model, and each user can rotate or revoke it at any time.

Q03

Which Weights & Biases actions can my agent take?

10 in total: 10 read, 0 write and 0 destructive. Your policies decide which of them each agent may call.

Q04

Can I make my agent read-only in Weights & Biases?

Yes. Allow read actions and deny writes in the Weights & Biases policy. Your agent can still look things up, and any write it attempts is blocked and logged.

Q05

Can my own backend call Weights & Biases too?

Yes. The same Weights & Biases connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.

Get started

Plug Weights & Biases into your agent.

Your users connect Weights & Biases once, under your brand. Your agent gets 10 actions behind your policies, with every call on the record.

Connect Weights & Biases Browse all apps →

Free to build · MCP + REST · Audit log on every plan