Enterprise model access and governance
Aggregated routing, unified control · OpenAI- and Anthropic-compatible. One endpoint, every major model.
01 · Access
One gateway,every major model
Point your base_url and API key at Tokenjoy and keep your OpenAI or Anthropic calls. What each model can do depends on the protocol and the upstream channel.
OpenAI · Anthropic · Streaming · Function calling · Vision
02 · Cost
Lower prices,billed per call
Selected models come with a platform discount, with prices listed in the pricing section below. The same model routes across channels by priority, and every call is logged with its tokens and cost, billed on real usage.
Platform discount · Per-call log · Billed on usage
03 · Governance
Governance,built in
Budgets cascade from company down to each member and stop calls once spent. Access is scoped by role and department. Every call and change is logged, filterable, and exportable.
Layered budgets · Role-based access · Audit trail
This month$23,545 / $30,000 limit
Running models across teams? Three things bite.
Costs you can't split, access you can't control, usage you can't see. One platform fixes all three.
Fragmented budget control
Management challengeShared credentials obscure cost ownership
Connect once, budget down to each person
One gateway for every model. Budgets flow from the company to departments, projects and members, and calls stop when a quota runs out.
- Hierarchical budget allocation
- Request quota checks
- Usage alerts and approvals
Disconnected access and audit management
Management challengePermissions and call records are difficult to correlate
Know who can use what, and who did
Choose which models each department and role can use. Every call and change is logged, so you can filter, export and look back whenever you need.
- Organization and role permissions
- Model access control
- Call and operation records
Limited operational visibility
Management challengeNo shared view of usage and service performance
See where the money goes, and if it's worth it
Break down cost trends and token usage by member, model and channel, and see cost and efficiency at a glance.
- Budget execution analysis
- Model and channel visibility
- Usage cost accounting
Core capabilities
At a glance
- 1Every app uses the same URL and key
- 2The gateway picks an available channel by priority
- 3Each request reaches the model it asked for
Show the full capability list
Model access and billing
4- Unified API access
- Centralize model calls through one gateway and API key system
- One endpoint
- APIs
- OpenAI- and Anthropic-compatible APIs; features depend on the model and channel
- Protocols
- Funds and quota management
- Use the enterprise wallet for settlement and budgets for quota allocation
- Separate management
- Usage billing
- Bill actual usage with token, usage-based and tiered pricing
- Actual usage
Models and routing
3- Channel routing
- Route model requests across channels by priority and weight
- Configurable
- Model capabilities
- Streaming, function calling and vision where supported by the model and channel
- Model-dependent
- Model access scope
- Set model access by department and API key, checked on each call
- Checked on calls
Service reliability
3Improve stability with channel failover, retries and rate limits.
- Failover
- Switch to another available channel on failure, as configured
- Channel failover
- Retries
- Retry upstream failures by policy and record status and errors
- As configured
- Rate limiting
- Set request rates and burst limits per API key
- Per key
Organization and permissions
3- Organization
- Manage departments and members with Feishu sync, CSV imports and invitations
- Team structure
- Roles
- Assign resource, budget and audit permissions by role
- Scoped permissions
- Approvals
- Approve quota increases, project budgets and member allocations
- As needed
Visibility and auditing
4- Call logs
- Review callers, models, usage and costs; export logs to CSV
- Scoped access
- Operation audit
- Trace actors, timestamps, objects and administrative changes
- Traceable changes
- Usage analysis
- Analyze usage and costs by department, member, model and channel
- Organization view
- Operational metrics
- Track requests, tokens, costs, latency and error rates
- Operational visibility
Role-based collaboration
Align permissions with responsibilities across management, development and auditing.
IT lead
Oversee model resources, budgets and governance
Company budgets · Models and channels · Access and audit
Team lead
Manage team quotas and budget execution
Dept and project quotas · Usage alerts · Quota approvals
Developer
Connect to models and review personal calls
Enterprise API keys · Compatible APIs · Personal call records
Ops and audit
Monitor operations and trace calls and changes
Call and operation audit · Metrics and usage · CSV exports
Every dollar, traced to a person and a model
Everyone's calls land in one pool, and the gateway splits the bill for you. Drill in level by level to see who used which model and how much.
Over budget? No big deal
One monthly quota per person; hit it and calls pause. Request a top-up, get a quick OK, and you're back.
Pay for what you use, viewable and exportable by department, member, model or channel.
Show the full check chain and billing rules
Hierarchical quota allocation
4- L1Company budgetPlan enterprise usage budgets
- L2Department / team budgetAllocate quota and aggregate usage
- L3Project / member quotaAssign quota by business need
- L4API keyLink quota, permissions and expiry
Request checks and accounting
8- 01KeyValidate the API key
- 02ModelIdentify the requested model
- 03CompanyCheck enterprise account status
- 04BalanceCheck wallet balance for platform channels; settle custom channels with suppliers
- 05Key stateCheck key status and expiry
- 06Model accessCheck department and key model access
- 07QuotaNo quotaCheck key quota and the relevant member, project or project-member allocation
- 08AccountingActual usageAccount for actual usage without duplicate entries
Read-only requests require identity and permission checks without consuming quota. In-flight calls settle on actual usage; final charges follow the bill.
Budget periods and alerts
Manage monthly budgets in the enterprise timezone, with usage alerts through configured channels.
Quota requests and approvals
Approve member and project quota increases within available allocations.
Model pricing
Explore base prices and default discounts. Enterprise rates, custom channel prices and final charges follow console settings and bills.
| Model | Input | Output | Cache |
|---|---|---|---|
| Deepseek | |||
| DeepSeek-V4-Flash-40%Chat | Time-based · 2 tiers | ||
| DeepSeek-V4-Pro-40%Chat | Time-based · 2 tiers | ||
| DeepSeek-V4.1-Flash-Economy-40%Chat | Time-based · 2 tiers | ||
| Doubao | |||
| 豆包 Seedance 2.0-8%Video | ≈¥0.41/秒 from≈¥0.44/秒 from | ||
| 豆包 Seedance 2.0 Fast-8%Video | ≈¥0.33/秒 from≈¥0.36/秒 from | ||
| 豆包 Seedance 2.5-8%Video | ≈¥0.62/秒 from≈¥0.67/秒 from | ||
| Moonshot | |||
| Kimi K3-30%Chat | ¥14¥20 | ¥70¥100 | read ¥1.4 / write ¥14read ¥2 / write ¥20 |
| Zhipu | |||
| GLM-5.2-30%Chat | ¥5.6¥8 | ¥19.6¥28 | read ¥1.4read ¥2 |
| GLM-5.3Chat | ¥8 | ¥28 | read ¥2 |
| GLM-5.3-FlashChat | ¥0.8 | ¥2.8 | read ¥0.23 |
CNY / 1M tokens; see model details for other billing units.
Flexible deployment, shared governance
Choose SaaS hosting or a private deployment to meet your operations and data management needs.
Request a product demo
Tell us what you need and we'll reach out within one business day to set up a demo. Or just open the demo environment and look around.
- Response time
- Contact within one business day
- Demo content
- Model access, permissions, budgets and call analysis
- Prep needed
- Just a browser, nothing to install
- Use of info
- Only for contact and demo scheduling
Focused on enterprise model governance
Connect model resources and enterprise applications to improve team collaboration and management visibility.
- Core team
- From Google, ByteDance, Baidu and Alipay, with experience in enterprise systems and product development.