Companies use external API calls to work with another company's product behind the scenes. Company A sends a request over the public internet to Company B, usually by HTTPS, carrying the details needed for the job and a key or token proving it has permission. Company B checks the request, does the work and sends back an answer.
Why bother with the extra work: a store shouldn't have to build a banking network, a helpdesk shouldn't have to train its own AI, and a small app shouldn't have to run a worldwide email service. API calls let Company A use Company B's specialist service inside its own product. Customers get one connected experience, while each company controls what information it shares and what actions it permits.
HTTPS protects the journey; API authorization controls access. Company A keeps private API keys on its server, shares only the information needed, and handles delays or errors if Company B cannot answer.
Processing Payments via Stripe 💳
The scenario: An online clothing store (Company A) needs to charge a customer for an order.
The API call: The store's checkout server sends Stripe (Company B) an encrypted request with the amount, currency and a secure payment reference — not its private API key in the customer's browser.
The action & response: Stripe works with the banking network and returns a payment status. Some payments need extra verification or finish later; Stripe can send a follow-up API message, called a webhook, when the status changes.
The result: Once payment is confirmed, the store marks the order paid and prepares it for shipping.
Signing In with Google 🔑
The scenario: You open a project management tool such as Notion (Company A) and choose “Sign in with Google” (Company B).
The user handshake: The tool sends your browser to Google's login page. You sign in directly with Google — the project management tool never sees your Google password.
The server-to-server API call: In a typical authorization-code sign-in flow, Google sends a temporary code back through your browser to the tool. The tool's server exchanges that code directly with Google's server over HTTPS, using its own registered credentials, such as a client secret, and any required security checks. The client secret identifies the app; it is not your password.
The action & response: Google validates the exchange and returns tokens: an ID token proving who signed in, and an access token for permitted API requests. Identity details such as a user ID, email and name can come from the verified ID token or a separate user-information API call, depending on the permissions granted.
The result: The tool checks the ID token and creates your signed-in session. This kind of shared sign-in is often called single sign-on; Google sign-in uses OpenID Connect, an identity layer built on OAuth 2.0.
Drafting Text with OpenAI 🤖
The scenario: A customer service tool (Company A) wants to draft a polite reply to a complaint.
The API call: Its server sends OpenAI (Company B) a prompt and the relevant complaint text, with unnecessary personal details removed.
The action & response: OpenAI returns generated text. The tool displays the draft for a customer service agent to review before sending it.
Calculating Delivery Routes with Google Maps 📍
The scenario: A food delivery app (Company A) wants to estimate when a driver will arrive.
The API call: Its server sends Google Maps (Company B) the driver's location and delivery destination to request a route and travel time.
The action & response: Google returns route details and an estimated duration. The app shows the route and arrival estimate; displaying the map itself may use separate browser or app requests for map images.
Sending Newsletters with SendGrid 📧
The scenario: A real estate platform (Company A) wants to send a newsletter to subscribers.
The API call: Its server sends SendGrid (Company B) the email content and intended recipients, splitting large mailings into batches within the service's limits.
The action & response: SendGrid acknowledges the request and handles delivery. Later event messages or reports tell the platform about deliveries, bounces, opens or clicks. Acceptance does not guarantee delivery or placement outside the spam folder.
Connecting Bank Accounts with Plaid 🏦
The scenario: A budgeting app (Company A) wants to let you connect a bank account with your permission.
The API call: Its server asks Plaid (Company B) for a temporary token to open a secure account-linking screen. After you authorize the connection, the app's server exchanges the returned reference for an access token.
The action & response: The app uses that token to request allowed information, such as balances or transactions. Plaid returns standardized data the budgeting app can show on your dashboard; available information depends on your consent, the bank and the service used.