Integration · Developer tools

Connect Claude & ChatGPT to Allowly

Ask your AI whether an agent may take an action under an Allowly authorization, which decisions were denied this week, or to revoke an authorization that is no longer needed. It can read policies, actions, authorizations and signed decision receipts, grant new authorizations, answer pending confirmations and resolve escalations. You manage agent permissions from a chat.

Connect Allowly free

Free plan, no credit card. Takes about a minute.

Try asking

“May the support agent issue a refund under authorization auth_812?”
The AI checks the action against the authorization and returns allow, deny, confirm or escalate, with a signed receipt.
“Which actions were denied in the last 7 days?”
It lists decision receipts with the decision deny in that time range, with action, resource and user.
“Grant user u_204 an authorization based on the read-only policy, expiring Friday.”
It finds the policy and creates the authorization for the user with the expiry you gave.

What your AI can do

Check actions

Check whether an agent may perform actions under an authorization, with resource, context and estimated cost, and get a decision with a signed receipt.

Authorizations and policies

List and read policies and registered actions, grant authorizations from a policy or explicit actions, and revoke them.

Confirmations, escalations and receipts

Approve or reject pending confirmations, resolve escalations, and list and read decision receipts by decision, action, user or time range.

“Approve the pending confirmation with nonce 7c1e.”
It answers the confirmation request with approved set to true.
“Revoke every active authorization for agent billing-bot.”
It lists the agent's active authorizations and revokes them. The agent can no longer act under them.
“Is our Allowly workspace fully set up?”
It reads the setup status of your workspace and shows what is configured.

33 tools for Allowly

These tools are switched on when you connect. You can switch any of them off, or require your approval before it runs.

answer_confirmationVisible to othersApprove or reject a pending confirmation request by its `nonce`, with `approved` true or false.
check_actionChanges dataCheck whether an AI agent may perform `actions` under an authorization (`authorization_id`), with optional `resource`, `context` and `estimated_cost_micros`. Returns allow, deny, confirm or escalate plus a signed receipt.
create_actionChanges dataCreate a new action. Required: name. API summary: Create Action.
create_authorizationChanges dataGrant an agent an authorization for a `user_id`, based on a `policy_id` or explicit `actions`, with optional budget limit, expiry and actions that need confirmation or escalation.
create_budget_settlementMoves moneyCreate a new budget settlement. Required: actual cost micros or check receipt id. Financial or irreversible action; may require approval. API summary: Settle Budget.
create_policyChanges dataCreate a new policy. Required: actions, agent id or policy id. API summary: Create Policy.
create_policy_create_policy_v1_setup_policies_postChanges dataCreate a new policy. Required: actions, agent id or policy id. API summary: Create Policy.
create_runtime_keyChanges dataCreate a new runtime key. API summary: Create Runtime Key.
create_sealChanges dataCreate a new seal. Required: profile, record sha256 or request id. API summary: Create Seal.
create_tombstoneChanges dataCreate a new tombstone. Required: resource pattern. API summary: Create Tombstone.
create_webhookDeletesCreate a new webhook. Destructive action; may require approval. API summary: Create Webhook Seal.
delete_actionDeletesPermanently delete a action by action id. Destructive action; may require approval. API summary: Delete Action.
delete_credentialsMoves moneyPermanently delete a credential. Financial or irreversible action; may require approval. API summary: Revoke Current Credential.
delete_policyDeletesPermanently delete a policy by policy id. Destructive action; may require approval. API summary: Delete Policy.
delete_policy_delete_policy_v1_setup_policies_policy_id_deleteDeletesPermanently delete a policy by policy id. Destructive action; may require approval. API summary: Delete Policy.
get_policyGet one policy by `policy_id`, including its allowed actions and confirmation and escalation rules.
get_policy_get_policy_v1_setup_policies_policy_id_getGet one policy by policy id. API summary: Get Policy.
get_receiptGet one signed decision receipt by `receipt_id`.
get_setup_statusGet the setup status of your Allowly workspace, showing what is configured.
get_shipmentGet one shipment by attempt id. API summary: Get Webhook Delivery.
get_webhook_receiptGet one receipt by receipt id. API summary: Get Webhook Receipt.
list_actionsList the actions registered in your workspace that policies can allow or restrict.
list_authorizationsList authorizations granted to agents, filtered by `user_id`, `agent_id` or `active_only`. Pages with `cursor`.
list_keysList keys. API summary: Get Webhook Keys.
list_policiesList policies that define which actions agents may take. Pages with `cursor`.
list_policies_list_policies_v1_setup_policies_getList policies. Optional filters: include archived. Results are paginated.
list_receiptsList decision receipts, filtered by `decision`, `action`, `resource`, `user_id`, `authorization_id` or a `from`/`to` time range. Pages with `cursor`.
list_tombstonesList tombstones. Results are paginated.
list_workspace_keysList keys for a workspace. API summary: Get Workspace Keys.
resolve_escalationChanges dataResolve an escalated action by `escalation_id` with a `resolution`, `resolved_by` and optional note.
revoke_authorizationDeletesRevoke an authorization by `authorization_id`, with optional `notes`. The agent can no longer act under it.
submit_policyVisible to othersPublish a policy. Has effects visible outside this system. API summary: Publish Policy.
update_actionChanges dataUpdate an existing action identified by action id. Only the fields you pass are changed. API summary: Update Action.

Set up in three steps

1
Pick the app
Create a free HeyMyApps account and choose Allowly from the app list.
2
Paste your key
Paste your Allowly runtime API key. In the Allowly dashboard (app.allowly.ai), create a runtime API key for your workspace and copy it. It starts with allowly_ and is shown only once.
3
Add the link to your AI
You get a personal MCP link. Add it to Claude, ChatGPT or Cursor:
1. Click your name, then Settings › Connectors › Add custom connector.
2. Paste your link as the Remote MCP server URL.
3. Switch it on from the + menu in a chat.

Questions

What can the AI do in Allowly?

It can check actions against an authorization, list and read policies, registered actions, authorizations and decision receipts, and read your workspace setup status. It can grant and revoke authorizations, approve or reject pending confirmations, and resolve escalations.

Does the AI see my Allowly credentials?

No. Your Allowly runtime API key is encrypted at rest and never shown to the AI. After you save it, it is not shown again, not even to you. The AI only sees the results of the tools it calls.

Can I control what the AI is allowed to do?

Yes. You choose which tools are switched on, so you can start with checks and lookups only. Revoking an authorization, answering a confirmation and granting new authorizations change what agents may do, and they can require your approval in HeyMyApps before they run. Every tool call is logged.

Does it work with ChatGPT?

Yes. In ChatGPT go to Settings › Apps & Connectors › Advanced and turn on Developer mode, then add your HeyMyApps link. Developer mode needs a paid ChatGPT plan: Plus, Pro, Business or Enterprise. The same link also works in Claude (Settings › Connectors › Add custom connector), Cursor and other MCP clients.

What does it cost?

HeyMyApps has a free plan with no credit card required. Paid plans bill per completed task. You also need your own Allowly account.

What is a signed receipt in Allowly?

Each time an action is checked, Allowly returns a decision (allow, deny, confirm or escalate) together with a signed receipt. The AI can list receipts by decision, action, resource, user, authorization or time range, and read any single receipt by its ID.