Zhipu AI open platform: the BigModel console overview

What the Zhipu AI open platform is

The Zhipu AI open platform is the developer console and management surface for the Z.ai ecosystem — not a separate product, but the operational backbone behind the chat interface and the API.

The Zhipu AI open platform is the BigModel developer console: the unified management surface for API keys, billing, project organisation, and usage analytics across the Z.ai product suite. It is not a model or a product that users interact with directly — it is the infrastructure layer that makes every other Z.ai surface manageable at scale. The chat Z AI surface, the Zhipuai API, and the Z AI chatbot integration patterns all draw on the same underlying inference infrastructure that the open platform exposes through the console.

The BigModel branding is the legacy name for the open platform that many engineers still know it by. The platform continues to operate under the BigModel brand in the developer-facing documentation and in the API base URL structure, even as the consumer-facing brand has transitioned to Z.ai. For practical purposes, BigModel and the Zhipu AI open platform refer to the same management console and the same API infrastructure. Teams that encounter the BigModel name in existing documentation, SDK examples, or forum answers are looking at the same platform covered here.

Console structure

The BigModel console is organised into four main sections — understanding the layout before logging in for the first time reduces orientation time.

The Zhipu AI open platform console is organised around four primary navigation sections. The Overview dashboard is the landing page after login — it shows a summary of recent API usage, active project count, current balance, and any platform announcements. This is the right place to check daily before starting any significant API work, as it surfaces rate limit status and upcoming billing cycle information without requiring drill-down into individual sections.

The API Keys section is where programmatic access credentials are created and managed. Keys can be labelled by project or environment (production, staging, test), enabling straightforward audit of which key was used for which request class. Revoking a compromised key from this section immediately invalidates it without affecting other keys on the account. The console shows the last four characters of each key and the creation date; it does not display the full key after initial creation, so copying it at creation time is mandatory.

The Projects section allows grouping API usage by project, with separate analytics and optional per-project spending caps. A team running three separate products against the Zhipuai API can create three projects and assign different API keys to each, then view the usage, cost, and error rate for each product independently in the analytics view. This is the feature that most teams only discover after they have already been billed for several months without the separation they needed.

The Billing section covers the account balance, top-up history, payment method management, and downloadable invoices. The platform operates on a pre-paid credit model for most international accounts — topping up a balance rather than committing to a periodic invoice gives teams better control over spending. Spend alerts can be configured to notify at a threshold before the balance runs out, which prevents unexpected API failures in production when a balance is depleted without warning.

API key management

The API key panel in the Zhipu AI open platform is the only place where valid keys for the Zhipuai API can be generated — there is no alternative path.

API key generation is exclusive to the Zhipu AI open platform console. There is no way to generate a valid Zhipuai API key through the chat interface, through the account settings page, or through any other surface. Accessing the API key panel requires a signed-in account — the Zhipuai login page is the entry point. After sign-in, the API Keys section is accessible from the primary navigation in the console.

Each key should be treated as a high-value credential from the moment it is generated. Store it in an environment variable or a secrets management system; never commit it to source code or paste it into a public forum. If a key is exposed, revoke it in the console immediately and generate a replacement. The time between exposure and revocation is the attack window — minimising it matters more than the rotation cadence for keys that have not been exposed.

Prompt history in the console

The prompt history feature in the Zhipu AI open platform gives a searchable log of recent API requests — useful for debugging and for understanding usage patterns across projects.

The Zhipu AI open platform console includes a prompt history section that logs recent API requests made against the account's keys. This is distinct from the conversation history in the chat interface — it is a raw request log that shows the full message array, the model selected, and the response received for each API call. The log is searchable by date range, by API key, and by project, which makes it a practical debugging tool when investigating unexpected responses or tracking down the source of a billing spike.

Prompt history retention is bounded by the account tier. Free tier accounts retain a shorter history window; paid tiers typically retain longer. The exact retention period is documented in the BigModel console under the prompt history section. For compliance-sensitive use cases where a longer audit log is required, the application layer should implement its own logging rather than relying solely on the console's prompt history.

Who the Zhipu AI open platform is for

The open platform serves a broad range of users from individual developers to enterprise teams — the feature set scales with usage complexity.

The Zhipu AI open platform is designed for anyone building on the GLM model family beyond casual chat use. At one end of the spectrum, an individual developer exploring the Zhipuai API for the first time will use the console primarily for API key generation and checking the free trial credit balance. At the other end, an enterprise team running multiple production applications against the API will use the full suite of project management, per-key analytics, spend alerts, and billing controls to manage the operational overhead of a multi-product AI integration portfolio.

The platform is also relevant for researchers and product teams who use the Z.ai chat interface as their primary surface but want visibility into their usage patterns and cost. The console's analytics view is useful even without any direct API integration — it provides context on what the chat interface is consuming that the interface itself does not surface. Research guidance from MIT's AI and Society group is useful background for teams evaluating governance and accountability frameworks for production AI platform usage.

Zhipu AI open platform features and use cases

Five key features of the Zhipu AI open platform with their primary use case and practical notes.

Practitioner note

"The project segmentation in the BigModel console was the feature that sold us on the Zhipu AI open platform for production use. Running five separate data pipeline integrations under one account with per-project billing breakdowns made the finance conversation straightforward — we could show exactly which pipeline was responsible for which cost."