Anthropic connector for AI agents

Tools your LLM ops agent reaches for on Anthropic, scoped per user.

CALL ANY TOOL

13 tools for Claude models in Anthropic: send messages, count tokens, look up models, create, track, and cancel message batches, and manage uploaded files.

anthropic_cancel_message_batch

Batch cancel message

Cancel a message batch that is still processing. Returns the batch object with processing_status canceling; requests already running may still finish and the rest end as canceled. Use to stop a batch you no longer need. Use anthropic_delete_message_batch to remove a finished batch.

Parameters

Name

Type

Required

Description

message_batch_id

string

Required

Id of the message batch, starting with msgbatch_. Returned by anthropic_create_message_batch and anthropic_list_message_batches.

anthropic_count_tokens

Count tokens

Count the input tokens a message request would use, without running the model. Returns input_tokens, the total across messages, system prompt and tools. Use before anthropic_create_message to check a prompt against context limits or estimate cost. It does not generate text and is not billed as a completion.

Parameters

Name

Type

Required

Description

messages

array

Required

Conversation turns, oldest first, alternating user and assistant roles. Each turn has a role (user or assistant) and content, which is either a plain string or an array of content blocks. Example: [{"role": "user", "content": "Hello, Claude"}]. A final assistant turn is treated as a prefill on models that allow it.

model

string

Required

Model id to run, for example claude-sonnet-5-5, claude-opus-5-5 or claude-haiku-5-5. Use anthropic_list_models to see the ids your key can use.

output_config

object

Optional

Output settings. Set effort to low, medium, high, xhigh or max to trade thoroughness for speed, and format to {"type": "json_schema", "schema": {...}} to request structured JSON output. Example: {"effort": "medium"}.

system

string

Optional

System prompt that sets context, role or instructions for the model. A plain string, or an array of text blocks such as [{"type": "text", "text": "You are a concise assistant."}].

thinking

object

Optional

Extended thinking configuration. Use {"type": "adaptive"} to let Claude decide when to think, {"type": "enabled", "budget_tokens": 2048} for a fixed thinking budget (at least 1024 and below max_tokens), or {"type": "disabled"}. Which types a model accepts depends on the model.

tool_choice

object

Optional

Controls how the model uses the supplied tools. Use {"type": "auto"} to let the model decide, {"type": "any"} to force some tool call, {"type": "tool", "name": "get_weather"} to force one tool, or {"type": "none"} to forbid tool calls.

tools

array

Optional

Client tool definitions the model may call. Each tool has a name, a description, and an input_schema that is a JSON Schema object. Example: [{"name": "get_weather", "description": "Get the weather for a city", "input_schema": {"type": "object", "properties": {"city": {"type": "string"}}, "required": ["city"]}}].

anthropic_create_message

Create message

Send a conversation to a Claude model and get the next assistant message, with optional system prompt, tools, thinking and structured output. Returns the message id, model, content blocks, stop_reason and token usage; non-streaming only. Use this for a single synchronous completion. Use anthropic_create_message_batch to run many requests asynchronously, and anthropic_count_tokens to size a prompt first. Charges apply to the API key's account for the tokens used.

Parameters

Name

Type

Required

Description

max_tokens

integer

Required

Maximum number of tokens to generate before stopping. The model may stop earlier. Example: 1024.

messages

array

Required

model

string

Required

Model id to run, for example claude-sonnet-5-5, claude-opus-5-5 or claude-haiku-5-5. Use anthropic_list_models to see the ids your key can use.

metadata_user_id

string

Optional

Opaque external id for the end user, such as a UUID or hash, used to help detect abuse. Do not use names or emails. Maximum 512 characters.

output_config

object

Optional

service_tier

string

Optional

Use auto to take priority capacity when your account has it, or standard_only to stay on standard capacity. One of: `auto`, `standard_only`.

stop_sequences

array

Optional

Custom strings that make the model stop generating when produced. Example: ["END"].

system

string

Optional

System prompt that sets context, role or instructions for the model. A plain string, or an array of text blocks such as [{"type": "text", "text": "You are a concise assistant."}].

temperature

number

Optional

Randomness from 0.0 to 1.0. Models released after Claude Opus 4.6 only accept 1.0 and reject other values with a 400 error, so leave this empty for them.

thinking

object

Optional

tool_choice

object

Optional

tools

array

Optional

top_k

integer

Optional

Only sample from the top K token options. Models released after Claude Opus 4.6 reject this parameter, so leave it empty for them.

top_p

number

Optional

Nucleus sampling cutoff from 0.0 to 1.0. Models released after Claude Opus 4.6 reject values below 0.99, so leave this empty for them.

anthropic_create_message_batch

Batch create message

anthropic_delete_file

Delete file

anthropic_delete_message_batch

Batch delete message

anthropic_get_file_metadata

Get file metadata

anthropic_get_message_batch

Batch get message

anthropic_get_message_batch_results

Get message batch results

anthropic_get_model

Get model

anthropic_list_files

List files

anthropic_list_message_batches

List message batches

anthropic_list_models

List models

Build your Agent

Same auth pattern across LangChain, OpenAI, Anthropic, and Google ADK.

import { ScalekitClient } from "@scalekit-sdk/node";
import { createReactAgent } from "@langchain/langgraph/prebuilt";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);

// Anthropic tools scoped to this user
const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["anthropic"], toolNames: [
    "anthropic_count_tokens",
    "anthropic_create_message",
    "anthropic_create_message_batch"] },
  pageSize: 100,
});

const agent = createReactAgent({ llm, tools });
await agent.invoke({ messages: [{ role: "user", content: "Count the tokens in this prompt, then send it to Claude" }] });

import OpenAI from "openai";
import { ScalekitClient } from "@scalekit-sdk/node";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);
const openai = new OpenAI();

const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["anthropic"] }, pageSize: 100,
});

const res = await openai.chat.completions.create({
  model: "gpt-5",
  messages: [{ role: "user", content: "Count the tokens in this prompt, then send it to Claude" }],
  tools,
});

// Execute the tool call with the user's vaulted Anthropic API key
await sk.tools.executeTool(res.choices[0].message.tool_calls[0], "user_123");

import Anthropic from "@anthropic-ai/sdk";
import { ScalekitClient } from "@scalekit-sdk/node";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);
const anthropic = new Anthropic();

const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["anthropic"] }, pageSize: 100,
});

const msg = await anthropic.messages.create({
  model: "claude-sonnet-5",
  max_tokens: 1024,
  messages: [{ role: "user", content: "Count the tokens in this prompt, then send it to Claude" }],
  tools,
});

// Tool call runs with the user's vaulted Anthropic API key
await sk.tools.executeTool(msg.content, "user_123");

import { Agent } from "@google/adk/agents";
import { ScalekitClient } from "@scalekit-sdk/node";

const sk = new ScalekitClient(env.SCALEKIT_ENV_URL, env.SCALEKIT_CLIENT_ID, env.SCALEKIT_CLIENT_SECRET);

const { tools } = await sk.tools.listScopedTools("user_123", {
  filter: { connectionNames: ["anthropic"] }, pageSize: 100,
});

const agent = new Agent({
  name: "llm_ops_agent",
  model: "gemini-2.5-pro",
  instruction: "Anthropic tools scoped to this user",
  tools,
});

await agent.run("Count the tokens in this prompt, then send it to Claude");

Try these prompts

Copy any prompt into your agent. Each maps directly to a Anthropic tool. Click to copy, paste into your agent, done.

Messages and tokens

Copy the prompt

Copied

Ask Claude to summarize this support thread in 3 bullet points.

Copy the prompt

Copied

Count the input tokens for this prompt and system message before I send it.

Copy the prompt

Copied

List the Claude models my API key can use, newest first.

Message batches

Copy the prompt

Copied

Submit these 500 classification prompts as one message batch.

Copy the prompt

Copied

Check the status of the batch I started this morning.

Copy the prompt

Copied

Fetch the results of yesterday's finished batch and show the errored requests.

Files

Copy the prompt

Copied

List the files uploaded to our workspace.

Copy the prompt

Copied

Show the metadata for the onboarding PDF and whether it is downloadable.

Copy the prompt

Copied

Delete the uploaded files we no longer need from last month's run.

SEE HOW AUTH WORKS

Each user connects their own Anthropic API key once, and Scalekit sends it with every call. Keys stay vaulted, every call is scope checked, and every action is logged.

Authorize

Your user connects

Anthropic

once. We tie it to their identity and the meetings they approved — no shared bot account, no org-wide access

Who:

user ‘A’

when:

Once per user

access:

Limited to user

Store

Their

Anthropic

token lives in a vault scoped to them. User A's meetings are never reachable by an agent acting for user B, even on the same connection

vault:

encrypted

scope:

per-user

tokens:

auto-refreshed

Resolve

When your agent calls a

Anthropic

tool, we fetch the right token server-side. It never touches your agent, never appears in the LLM context, never shows up in your logs

speed:

~40ms

check:

before every call

seen by:

nobody

Audit

Every

Anthropic

tool call is logged — who triggered it, which meeting was fetched, what came back. 90 days of history, tied to the user who authorized it

history:

90 days

export:

SIEM-ready

logged:

every call

Test other agents

See the same per-user auth pattern across other AI and LLM connectors.

GTM and RevOps Teams

Competitive intelligence briefing agent

Scans Gong calls for competitor mentions, matches each one to its Notion battlecard, and DMs every affected rep a single Slack digest per cycle. Every call runs as the PMM who owns the briefing, never a shared bot.

Engineering Teams

Slack triage

Polls Slack for new messages, classifies bugs and support requests with a LangGraph router, files GitHub issues or Zendesk tickets, and confirms in the thread.

Why Scalekit

Secure your agent's access. Connectors ship in minutes

01.

Shared tokens break per-user analytics

A shared Anthropic API key looks fine in a demo. In production every message batch looks like one service account, and you cannot tell which user triggered it. Scalekit resolves the credential of the actual user who triggered the agent, never a shared bot.

// shared key
audit → bot_service_account

// scalekit
audit → user_abc ✓

02.

Authentication is not authorization

03.

Multi-tenancy is architectural

04.

Anthropic today. Ten connectors tomorrow.

“Our agents act across Salesforce, Gong, Google Drive, and more, on behalf of every customer. Scalekit behind the scenes meant we can keep adding tools without ever rebuilding how credentials or tool calling work.”

Venu Madhav Kattagoni

Head of Engineering / Von

FAQs

Frequently Asked Questions

Does the agent access Anthropic as the user or through a shared API key?

As the user. Each user connects their own Anthropic API key once, and Scalekit sends it with every call. Audit logs attribute every message and batch to that user, not a shared service account.

Where is the Anthropic API key stored?

In Scalekit's managed AES-256 token vault, namespaced per tenant. To rotate it, the user reconnects with a new API key. Revocation is a single dashboard action. Credentials never appear in prompts, logs, or LLM context.

Can I limit what the agent does in Anthropic?

Yes. Filter by tool name in listScopedTools to expose only what you want, for example messages, token counting, and batch status without anthropic_delete_file, which permanently deletes an uploaded file, or anthropic_delete_message_batch. Scalekit also enforces scope checks before every API call.

What happens when a user revokes Anthropic access?

The connection is invalidated on the next tool call. Subsequent requests for that user fail closed with a clear error. Other users in the tenant remain unaffected. The event is logged for audit.

Which Claude models can the agent call?

The models available to the user's Anthropic API key. anthropic_list_models returns them newest first with token limits and capabilities, and anthropic_get_model looks one up by id or alias, including its deprecation and retirement dates.