Bland AI
Send AI phone calls, manage conversational pathways, phone numbers, and voices with Bland.
About
Bland AI ships in the w6w first-party pack. It declares 19 actions, 2 health checks, and the host runs its code in a sandbox that never sees the credential.
- App id
io.w6w.blandai- Version
- 0.1.0
- Author
- w6w
- Licence
- MIT
- Categories
- AI & Machine Learning · Communication
Overview
Bland is an enterprise voice AI platform: telephony, speech-to-text, language model, text-to-speech, and conversational-pathway orchestration in one stack. This app covers Bland’s core call surface — dispatch a phone call with a task prompt or a conversational pathway, list and inspect calls (including the full transcript), stop an active call or all of them, transfer a live call to another number, and grade a completed call against a goal and a set of questions with AI.
Beyond calls, it covers the two other things most Bland automations need: conversational pathways (list, get, create, update, and delete the node/edge graphs an agent follows) and the account’s phone numbers and voices (list and inspect inbound numbers, purchase a new number, and list or inspect the voices — curated, cloned, or library — available to speak on a call).
Health checks track Bland’s own status page (status.bland.ai) alongside the account’s live credit balance and status, so a workflow can see both whether Bland itself is healthy and whether this account still has call credit left.
Bland’s reference also documents a much larger surface this v1 does not cover — agent testing and evals, knowledge bases, widgets, SMS/RCS/iMessage messaging, SIP trunks, memory, triage, custom dialing pools, and personas — left out deliberately rather than built against guesses; see this app’s README for what was left out and why.
Build with Bland AI
Three routes to the same 19 actions. The Workflow tab is generated from Bland AI's own manifest and carries its real ids, so it is copy-pasteable; the Code and CLI examples are the same call for any action on any app, so every app-specific value in them is a blank you fill in.
Get Account
account-get Fetch the connected account's status, billing balance, and total call count.
Analyze Call
call-analyze Analyze a call's transcript against a goal and a set of questions using AI.
List Active Calls
call-list-active Retrieve all currently queued or in-progress calls for this account.
List Numbers
number-list Retrieve every inbound phone number configured for this account, with settings.
Purchase Phone Number
number-purchase Purchase a new phone number ($15/mo, billed to the account's stored payment method).
Create Pathway
pathway-create Create a new conversational pathway. Use pathway-update to add nodes and edges.
List Pathways
pathway-list Returns every conversational pathway in the account, including nodes and edges.
A workflow step names the app and the action, and the editor fills in the
connection when you pick one. This is the Step shape from the
workflow spec, carrying Bland AI's real ids.
{
"manifestVersion": "2",
"name": "blandai-example",
"steps": [
{
"id": "call-analyze",
"uses": {
"app": "io.w6w.blandai",
"action": "call-analyze",
"connection": "conn_YOUR_CONNECTION_ID"
},
"with": {
"callId": "<callId>",
"goal": "<goal>",
"questions": "<questions>"
}
}
]
}Here are some of the things you can do
-
Analyze Call
performcall-analyze -
Get Account
readaccount-get -
Get Call
readcall-get -
List Calls
readcall-list -
List Active Calls
readcall-list-active
+14 more actions available
Every app-specific value here is a blank you have to fill in. An
app action is reached through the connection that authenticates it, so the
address is a connection id, not the app id — and connections belong to your account,
so a public page cannot know yours. Create one for Bland AI, then fill in
the three blanks: conn_YOUR_CONNECTION_ID, the action key, and the
parameters that action declares. The call itself is real — the shape is transcribed
from the studio's own snippet builder, which prints the same kind of blanks — but
nothing in it is specific to Bland AI. The Workflow tab is where this app's
real ids are.
npm install @w6w/sdkyarn add @w6w/sdkpnpm add @w6w/sdkdeno add npm:@w6w/sdkimport { W6wClient, isActionRun } from "@w6w/sdk";
// Reads W6W_BASE_URL and W6W_TOKEN from the environment when omitted.
const client = new W6wClient();
const envelope = await client.run({
urn: "conn_YOUR_CONNECTION_ID",
action: "call-analyze",
payload: {
callId: "<value>",
goal: "<value>",
questions: "<value>",
},
});
if (isActionRun(envelope)) console.log(envelope.value);npm install -g @w6w/cliw6w run conn_YOUR_CONNECTION_ID --action call-analyze --payload '{"callId":"<value>","goal":"<value>","questions":"<value>"}'Give an AI agent Bland AI — without giving it Bland AI's credentials. One MCP endpoint exposes every app, function and workflow the caller is entitled to, as tools it can discover and run. Access is granted per team while we onboard.
One tool call{
"name": "w6w_invoke",
"arguments": {
"ref": "app:io.w6w.blandai#call-analyze",
"input": {
"callId": "<callId>",
"goal": "<goal>",
"questions": "<questions>"
}
}
}
Every tool names its target with a single ref. The
app: form above doesn't name a connection at all — the
host resolves which of the caller's Bland AI connections to sign
with, and refuses rather than guesses when the answer is ambiguous.
What the agent gets
Credentials it can't read
The token is attached host-side, at the moment of the call. It is never a tool argument, never in the model's context, and never in a transcript — so a prompt injection has nothing to exfiltrate.
A tool surface scoped to the caller
Tools are derived per end user from what that person has actually connected and is entitled to — not one shared bot identity carrying the union of everyone's access.
A durable workflow in one call
Multi-step work runs on the workflow engine and returns a run handle the agent can poll — retries, branching and state survive the conversation that started them.
Health-aware discovery
Bland AI's declared health checks are on the surface too, so an agent can tell "the vendor is down" from "your credential expired" before it burns a retry on either.
The MCP surface is part of the hosted platform. Bland AI itself is MIT, and the runtime that executes it is source-available (FSL).
Health checks
Bland AI declares its own checks, so its health is a property of the app rather than something the host guesses at.
Bland platform status
Component status from status.bland.ai, across every regional group Bland reports.
Credit balance
Remaining pay-as-you-go call credit and account status, read from GET /v1/me.