Skip to content

Connectors · Developer & infrastructure

Hyperbrowser MCP server for AI agents.

actionlake is an MCP gateway: one remote MCP server that connects Claude and other MCP clients to Hyperbrowser and the rest of your systems. It gives your agent 105 Hyperbrowser actions (52 reads, 34 writes and 19 that wait for a confirmation) and holds the Hyperbrowser credential itself, so the agent never does. It paces Hyperbrowser's limits and records every call.

100,000 free calls a month. No card, no email, no password: press Start a free sandbox on the next screen and keep the key it shows you.

claude mcp add --transport http actionlake https://connect.actionlake.com/mcp
the first call
do {  action: "hyperbrowser.profiles.list",  intent: "check which Hyperbrowser account is connected"}# Get list of profiles · read · 1 upstream call
  • 105actionsin one signed file
  • 52readlook, never change
  • 34writecreate and update
  • 19wait for a yesdestructive, confirmed first

Set up

Connect Hyperbrowser in three steps.

  1. 01

    Add actionlake to Claude.

    Run the command once in Claude Code; its consent page can create a free sandbox: 100,000 calls a month, no card. In the Claude app, add a custom connector with the URL below and paste your key on the consent page. Creating a new sandbox from the Claude app is not yet tested by us.

    https://connect.actionlake.com/mcp
  2. 02

    Connect Hyperbrowser.

    API key. No consent screen: your agent can connect it for you, but the key then passes through your chat. To keep it out of the model's context, paste it in the console instead. actionlake stores what you paste write-only, sealed under your tenant's key.

    • api_keyHyperbrowser API key: in the Hyperbrowser dashboard open Settings > API Keys and create a key
    account { op: "connections.requirements", service: "hyperbrowser" }
    # free: exactly what to paste, and what is found for you
  3. 03

    Make the first call. Check before a write.

    Ask in plain words; Claude picks the action. Before anything changes, check validates the call and prices it, for free.

    do    { action: "hyperbrowser.profiles.list" }# get list of profiles · read · 1 upstream callcheck { action: "hyperbrowser.profiles.create", input: { … } }# free: inputs, permissions, cost and confirmations, nothing sent

Hyperbrowser at a glance

Sign-in
API key. No consent screen: your agent can connect it for you, but the key then passes through your chat. To keep it out of the model's context, paste it in the console instead.
You provide
api_key
Limits we pace to
60 calls a minute, per connection
API covered
Hyperbrowser API 1.0.0 (OpenAPI 3.0.1)api.hyperbrowser.ai
Built from
Hyperbrowser's documentation, read 23 October 2026docs.hyperbrowser.aidocs.hyperbrowser.ai
Service file
Version 1.0.0, signed (Ed25519)
Cached reads
15 of 52 reads, for up to 5 minutes; a write clears what it affects
Shared entities
1, so ask can line Hyperbrowser up with your other systems
Data
Processed and stored in Sydney, Australia. No US region yet

Capabilities

What an AI agent can do in Hyperbrowser.

105 actions in 25 areas. Every one carries its effect, so a key can allow the reads and hold the rest.

  • 52readlook, never change
  • 34writecreate and update
  • 19destructivewait for a confirmation
  1. Sessions

    10

    5 read · 4 write · 1 destructive

    Create new session · Run manual CAPTCHA evaluation · Get list of sessions · Fill a credential

  2. Sandboxes

    9

    2 read · 6 write · 1 destructive

    Create new sandbox · Create a sandbox memory snapshot · Get list of sandboxes · Start a sandbox process with secrets

  3. Secrets

    8

    3 read · 3 write · 2 destructive

    Add a secret version · Create a secret · List secrets · Get a secret

  4. Image builds

    5

    2 read · 2 write · 1 destructive

    Create an image build · Get list of image builds · Complete an image build upload · Get image build by ID

  5. Vaults

    5

    2 read · 2 write · 1 destructive

    Create a vault · List vaults · Get a vault · Update a vault

  6. Webmcp

    5

    2 read · 2 write · 1 destructive

    Invoke a WebMCP tool · List WebMCP tools · Start a WebMCP invocation · Get a WebMCP invocation result

  7. Browser use tasks

    4

    2 read · 1 write · 1 destructive

    Start a browser use task · Get browser use task status and results · Get browser use task status · Stop a browser use task

  8. Claude computer use tasks

    4

    2 read · 1 write · 1 destructive

    Start a claude computer use task · Get claude computer use task status and results · Get claude computer use task status · Stop a claude computer use task

17 more areas · 55 actions
  • Gemini computer use tasks 4
  • Grok computer use tasks 4
  • Hyper agent tasks 4
  • Jev tasks 4
  • Meta computer use tasks 4
  • Openai cua tasks 4
  • Profiles 4
  • Volumes 4
  • Batch scrape jobs 3
  • Crawl jobs 3
  • Extract jobs 3
  • Scrape jobs 3
  • Snapshots 3
  • Web crawl jobs 3
  • Images 2
  • Web 2
  • Extensions 1

Top actions

The Hyperbrowser API, shaped for agents.

The ids your agent types, what each does, and the one name it shares with other systems.

ActionWhat it doesEffectSame name in
hyperbrowser.web.fetchFetch a web pageFetch a web page.readthis connector
hyperbrowser.web.searchSearch the webSearch the web.readthis connector
hyperbrowser.volumes.listGet list of volumesGet list of volumes.readthis connector
hyperbrowser.profiles.listGet list of profilesGet list of profiles.readthis connector
hyperbrowser.profiles.createCreates a new profileCreates a new profile.writethis connector
hyperbrowser.sessions.createCreate new sessionCreate new session.writethis connector
hyperbrowser.crawl_jobs.startStart a crawl jobStart a crawl job.writethis connector
hyperbrowser.scrape_jobs.startCreate new scrape jobCreate new scrape job.writethis connector
hyperbrowser.extract_jobs.startStart an extract jobStart an extract job.writethis connector
hyperbrowser.browser_use_tasks.startStart a browser use taskStart a browser use task.writethis connector
hyperbrowser.hyper_agent_tasks.startStart a HyperAgent taskStart a HyperAgent task.writethis connector
hyperbrowser.profiles.deleteDelete profile by IDDelete profile by ID.destructivethis connector

Try it

Prompts to paste into Claude.

Each one names the Hyperbrowser actions it will call, so you know what it reads and what it changes before you send it.

  1. Readread

    Get list of profiles in Hyperbrowser and tell me what stands out.

    callshyperbrowser.profiles.list

  2. Writewrite

    Creates a new profile in Hyperbrowser. Show me exactly what you will send before you save it.

    callshyperbrowser.profiles.create

Governance for Hyperbrowser

Destructive actions wait for a yes.

19 Hyperbrowser actions are destructive. These are all of them:

  • Delete profile by IDhyperbrowser.profiles.delete
  • Stop a sessionhyperbrowser.sessions.stop
  • Stop a browser use taskhyperbrowser.browser_use_tasks.stop
  • Stop a HyperAgent taskhyperbrowser.hyper_agent_tasks.stop
  • Stop a claude computer use taskhyperbrowser.claude_computer_use_tasks.stop
  • Delete sandbox imagehyperbrowser.images.delete
  • Delete a vaulthyperbrowser.vaults.delete
  • Delete a secrethyperbrowser.secrets.delete
  • Delete a volumehyperbrowser.volumes.delete
  • Delete sandbox snapshothyperbrowser.snapshots.delete
  • Delete a secret versionhyperbrowser.secrets.delete_version
  • Stop a jev computer use taskhyperbrowser.jev_tasks.stop
  • Stop a sandboxhyperbrowser.sandboxes.stop
  • Cancel an image buildhyperbrowser.image_builds.cancel
  • Stop a CUA taskhyperbrowser.openai_cua_tasks.stop
  • Cancel a WebMCP invocationhyperbrowser.webmcp.cancel_invocation
  • Stop a grok computer use taskhyperbrowser.grok_computer_use_tasks.stop
  • Stop a meta computer use taskhyperbrowser.meta_computer_use_tasks.stop
  • Stop a gemini computer use taskhyperbrowser.gemini_computer_use_tasks.stop

A destructive action runs only with confirm: true. Under the human policy the agent cannot confirm for itself: the run stops until a named person approves, in the console or from a signed link in an email. The cap on a key is refused before the call is sent, never after.

dohyperbrowser.profiles.deleteconfirm_policy: human

Delete profile by ID?

intent: the user asked me to delete profile by ID

run suspended · waiting for a named person · key hyperbrowser-agent cannot confirm itself
In the console, or from an emailed signed link.
a key for this connector
account {  op: "keys.create", name: "hyperbrowser-agent",  allowed_actions: ["hyperbrowser.*"], cap_calls: 2000,  confirm_policy: "human_for_writes",  approval_rules: [    { match: "hyperbrowser.profiles.delete", approver: "[email protected]" },    { match: "hyperbrowser.sessions.stop", approver: "[email protected]" },    { match: "hyperbrowser.browser_use_tasks.stop", approver: "[email protected]" }  ]}
Illustrative. Scoped to Hyperbrowser only, capped at 2,000 upstream calls, every write confirmed, and the destructive ones approved by a named person.

Cost

What Hyperbrowser costs on actionlake.

Every Hyperbrowser action is one upstream call, the only thing actionlake bills. Searching, check, cost, account and any call Hyperbrowser refused with 401, 403, 429 or a 5xx are free. Paid plans start at US$49 a month for 250,000 calls, with overage from US$0.30 down to US$0.18 per 1,000.

  • 1upstream call per actionall 105 Hyperbrowser actions
  • 105actions pageeach page of a long list is one call
  • 15reads can be cacheda cache hit costs nothing
  • 100,000calls a month, freepaid plans from US$49 a month for 250,000

See pricing

Action index

Every Hyperbrowser action.

105 actions by area: the id your agent calls, what it does, and its effect.

Sessions10
  • hyperbrowser.sessions.createCreate new sessionwrite
  • hyperbrowser.sessions.evaluate_captchaRun manual CAPTCHA evaluationwrite
  • hyperbrowser.sessions.fill_credentialsFill a credentialwrite
  • hyperbrowser.sessions.getGet session by IDread
  • hyperbrowser.sessions.get_downloads_urlGet session downloads URLread
  • hyperbrowser.sessions.get_recording_urlGet session recording URLread
  • hyperbrowser.sessions.get_video_recording_urlGet session video recording URLread
  • hyperbrowser.sessions.listGet list of sessionsread
  • hyperbrowser.sessions.stopStop a sessiondestructive
  • hyperbrowser.sessions.updateUpdate a running sessionwrite
Sandboxes9
  • hyperbrowser.sandboxes.createCreate new sandboxwrite
  • hyperbrowser.sandboxes.create_snapshotCreate a sandbox memory snapshotwrite
  • hyperbrowser.sandboxes.expose_portExpose a sandbox portwrite
  • hyperbrowser.sandboxes.getGet sandbox by IDread
  • hyperbrowser.sandboxes.listGet list of sandboxesread
  • hyperbrowser.sandboxes.start_processStart a sandbox process with secretswrite
  • hyperbrowser.sandboxes.stopStop a sandboxdestructive
  • hyperbrowser.sandboxes.unexpose_portRemove an exposed sandbox portwrite
  • hyperbrowser.sandboxes.update_networkUpdate sandbox network policywrite
Secrets8
  • hyperbrowser.secrets.add_versionAdd a secret versionwrite
  • hyperbrowser.secrets.createCreate a secretwrite
  • hyperbrowser.secrets.deleteDelete a secretdestructive
  • hyperbrowser.secrets.delete_versionDelete a secret versiondestructive
  • hyperbrowser.secrets.getGet a secretread
  • hyperbrowser.secrets.get_public_keyGet the secrets public keyread
  • hyperbrowser.secrets.listList secretsread
  • hyperbrowser.secrets.updateUpdate a secretwrite
Image builds5
  • hyperbrowser.image_builds.cancelCancel an image builddestructive
  • hyperbrowser.image_builds.completeComplete an image build uploadwrite
  • hyperbrowser.image_builds.createCreate an image buildwrite
  • hyperbrowser.image_builds.getGet image build by IDread
  • hyperbrowser.image_builds.listGet list of image buildsread
Vaults5
  • hyperbrowser.vaults.createCreate a vaultwrite
  • hyperbrowser.vaults.deleteDelete a vaultdestructive
  • hyperbrowser.vaults.getGet a vaultread
  • hyperbrowser.vaults.listList vaultsread
  • hyperbrowser.vaults.updateUpdate a vaultwrite
Webmcp5
  • hyperbrowser.webmcp.cancel_invocationCancel a WebMCP invocationdestructive
  • hyperbrowser.webmcp.get_invocationGet a WebMCP invocation resultread
  • hyperbrowser.webmcp.invokeInvoke a WebMCP toolwrite
  • hyperbrowser.webmcp.list_toolsList WebMCP toolsread
  • hyperbrowser.webmcp.start_invocationStart a WebMCP invocationwrite
Browser use tasks4
  • hyperbrowser.browser_use_tasks.getGet browser use task status and resultsread
  • hyperbrowser.browser_use_tasks.get_statusGet browser use task statusread
  • hyperbrowser.browser_use_tasks.startStart a browser use taskwrite
  • hyperbrowser.browser_use_tasks.stopStop a browser use taskdestructive
Claude computer use tasks4
  • hyperbrowser.claude_computer_use_tasks.getGet claude computer use task status and resultsread
  • hyperbrowser.claude_computer_use_tasks.get_statusGet claude computer use task statusread
  • hyperbrowser.claude_computer_use_tasks.startStart a claude computer use taskwrite
  • hyperbrowser.claude_computer_use_tasks.stopStop a claude computer use taskdestructive
Gemini computer use tasks4
  • hyperbrowser.gemini_computer_use_tasks.getGet gemini computer use task status and resultsread
  • hyperbrowser.gemini_computer_use_tasks.get_statusGet gemini computer use task statusread
  • hyperbrowser.gemini_computer_use_tasks.startStart a gemini computer use taskwrite
  • hyperbrowser.gemini_computer_use_tasks.stopStop a gemini computer use taskdestructive
Grok computer use tasks4
  • hyperbrowser.grok_computer_use_tasks.getGet grok computer use task status and resultsread
  • hyperbrowser.grok_computer_use_tasks.get_statusGet grok computer use task statusread
  • hyperbrowser.grok_computer_use_tasks.startStart a grok computer use taskwrite
  • hyperbrowser.grok_computer_use_tasks.stopStop a grok computer use taskdestructive
Hyper agent tasks4
  • hyperbrowser.hyper_agent_tasks.getGet HyperAgent task status and resultsread
  • hyperbrowser.hyper_agent_tasks.get_statusGet HyperAgent task statusread
  • hyperbrowser.hyper_agent_tasks.startStart a HyperAgent taskwrite
  • hyperbrowser.hyper_agent_tasks.stopStop a HyperAgent taskdestructive
Jev tasks4
  • hyperbrowser.jev_tasks.getGet jev computer use task status and resultsread
  • hyperbrowser.jev_tasks.get_statusGet jev computer use task statusread
  • hyperbrowser.jev_tasks.startStart a jev computer use taskwrite
  • hyperbrowser.jev_tasks.stopStop a jev computer use taskdestructive
Meta computer use tasks4
  • hyperbrowser.meta_computer_use_tasks.getGet meta computer use task status and resultsread
  • hyperbrowser.meta_computer_use_tasks.get_statusGet meta computer use task statusread
  • hyperbrowser.meta_computer_use_tasks.startStart a meta computer use taskwrite
  • hyperbrowser.meta_computer_use_tasks.stopStop a meta computer use taskdestructive
Openai cua tasks4
  • hyperbrowser.openai_cua_tasks.getGet CUA task status and resultsread
  • hyperbrowser.openai_cua_tasks.get_statusGet CUA task statusread
  • hyperbrowser.openai_cua_tasks.startStart a CUA taskwrite
  • hyperbrowser.openai_cua_tasks.stopStop a CUA taskdestructive
Profiles4
  • hyperbrowser.profiles.createCreates a new profilewrite
  • hyperbrowser.profiles.deleteDelete profile by IDdestructive
  • hyperbrowser.profiles.getGet profile by IDread
  • hyperbrowser.profiles.listGet list of profilesread
Volumes4
  • hyperbrowser.volumes.createCreate a volumewrite
  • hyperbrowser.volumes.deleteDelete a volumedestructive
  • hyperbrowser.volumes.getGet volume by IDread
  • hyperbrowser.volumes.listGet list of volumesread
Batch scrape jobs3
  • hyperbrowser.batch_scrape_jobs.getGet batch scrape job status and resultsread
  • hyperbrowser.batch_scrape_jobs.get_statusGet batch scrape job statusread
  • hyperbrowser.batch_scrape_jobs.startStart a batch scrape jobwrite
Crawl jobs3
  • hyperbrowser.crawl_jobs.getGet crawl job status and resultsread
  • hyperbrowser.crawl_jobs.get_statusGet crawl job statusread
  • hyperbrowser.crawl_jobs.startStart a crawl jobwrite
Extract jobs3
  • hyperbrowser.extract_jobs.getGet extract job status and resultsread
  • hyperbrowser.extract_jobs.get_statusGet extract job statusread
  • hyperbrowser.extract_jobs.startStart an extract jobwrite
Scrape jobs3
  • hyperbrowser.scrape_jobs.getGet scrape job status and resultread
  • hyperbrowser.scrape_jobs.get_statusGet scrape job statusread
  • hyperbrowser.scrape_jobs.startCreate new scrape jobwrite
Snapshots3
  • hyperbrowser.snapshots.deleteDelete sandbox snapshotdestructive
  • hyperbrowser.snapshots.getGet sandbox snapshotread
  • hyperbrowser.snapshots.listGet list of sandbox snapshotsread
Web crawl jobs3
  • hyperbrowser.web_crawl_jobs.getGet web crawl job resultsread
  • hyperbrowser.web_crawl_jobs.get_statusGet web crawl job statusread
  • hyperbrowser.web_crawl_jobs.startStart a web crawl jobwrite
Images2
  • hyperbrowser.images.deleteDelete sandbox imagedestructive
  • hyperbrowser.images.listGet list of sandbox imagesread
Web2
  • hyperbrowser.web.fetchFetch a web pageread
  • hyperbrowser.web.searchSearch the webread
Extensions1
  • hyperbrowser.extensions.listList all extensionsread

Questions

The Hyperbrowser MCP server, answered.

Is there a Hyperbrowser MCP server for Claude?

Yes. actionlake is an MCP gateway: a remote MCP server that gives Claude, or another MCP client, 105 Hyperbrowser actions across sessions, sandboxes, secrets and image builds, through the same eight tools it uses for every other connector. actionlake is independent; it is not made or endorsed by Hyperbrowser.

What do I need to connect Hyperbrowser?

Hyperbrowser API key: in the Hyperbrowser dashboard open Settings > API Keys and create a key. The free call account { op: "connections.requirements", service: "hyperbrowser" } lists exactly what to paste.

Can an AI agent delete or change things in Hyperbrowser?

Only if its key allows it, and only with an explicit confirmation. 19 Hyperbrowser actions are destructive, including delete profile by ID, stop a session and stop a browser use task. Each runs only with confirm: true. Under the human confirm policy, or an approval rule on the key, the run waits until a named person approves it.

Will an AI agent use up my Hyperbrowser API limit?

actionlake paces every call to the limits in the Hyperbrowser service file: 60 calls a minute per connection. Repeated reads are answered from cache (15 Hyperbrowser actions can be), which never reaches Hyperbrowser.

What does a Hyperbrowser action cost?

One upstream call, the only unit actionlake bills. A list that pages bills one call per page (105 Hyperbrowser actions page). Cache hits, searches, dry runs and calls Hyperbrowser refused with 401, 403, 429 or a 5xx are free. The Free plan includes 100,000 calls a month with no card. Paid plans start at US$49 a month for 250,000 calls, with overage from US$0.30 down to US$0.18 per 1,000.

Which Hyperbrowser API does it cover?

The Hyperbrowser API 1.0.0 (OpenAPI 3.0.1) at api.hyperbrowser.ai. It is built from Hyperbrowser's documentation as read on 23 October 2026. 105 actions in 25 areas, each one in the signed service file (version 1.0.0).

Does actionlake keep my Hyperbrowser data, and where?

actionlake processes and stores data in Sydney, Australia; there is no US region yet. Reads may be cached for up to 5 minutes, and a write clears the reads it affects. Each run's request and response is kept, sealed, for your plan's trace window: 1 day on Free, 3 days on Starter, 7 days on Pro, 14 days on Team, 30 days on Business and 30 days on Enterprise. Your Hyperbrowser credential is sealed under your tenant's key, never logged and never returned. Your traffic is not used to train models.

Which AI clients can use it?

Claude Code, tested by us end to end, and the Claude app as a custom connector, tested by us with an existing key. Other MCP clients that support remote servers over Streamable HTTP (MCP revisions 2025-06-18 and 2025-11-25) and OAuth dynamic client registration, or a bearer header, should connect the same way; we have not tested them. Or plain REST, under the same action names.

Give your agent Hyperbrowser today.

Start free - 100,000 calls/month, no card. No signup. Hard caps on every key. Cached answers are free.