One endpoint for 200+ models worldwide
Every AI dollar accounted for
Text, multimodal, structured decision, image, video, speech and embeddings — seven model types your engineers call through one API, while IT and finance see every cost by department and business task.
Unified access
OpenAI-compatible. Change one line — the base URL — and your existing code can call 200+ models worldwide, with no SDK to maintain per vendor.
Cost visibility
Every call is attributed to a department and a business task. Department bills, cost per task and monthly chargeback reports are always available to view and export.
Reliable and controlled
When an upstream provider rate-limits or times out, calls retry automatically or fail over to a backup model. Department budgets have hard caps and early alerts, so there are no surprise bills.
One API, 200+ models worldwide
Text, multimodal, structured decision (such as TypeSafe Jev), image, video, speech and embedding models from leading vendors worldwide. Here are a few popular flagships — pick one to see ready-to-run code.
Break the AI bill down by department and task
No more single lump-sum invoice. IT can answer three questions: which department spent it, on which tasks, and what next month's budget should be.
| Department | Cost center | Main model | Spend this month | Monthly budget | Used | Status |
|---|---|---|---|---|---|---|
| Core Engineering | CC-RD-01 | Claude Sonnet 5.5 | $9,412 | $11,000 | 85.6% | Near limit |
| Customer Service | CC-CS-02 | GPT-5.6 Sol | $5,834 | $8,300 | 70.3% | On track |
| Legal & Compliance | CC-LG-03 | GPT-6 | $3,454 | $4,160 | 83.0% | Near limit |
| Business Analytics | CC-BI-04 | Gemini 4 Argon | $2,104 | $3,440 | 61.2% | On track |
Sample data · October 2026 · Each department has its own Virtual Key; the owner is notified when spend passes 80% of budget
Sample data · Average cost per task = total task spend ÷ number of calls; tasks are tagged with the x-unirouter-task header
| Department | Claude Opus 5.5 | Claude Sonnet 5.5 | GPT-6 | GPT-5.6 Sol | Gemini 4 Argon | Total |
|---|---|---|---|---|---|---|
| Core Engineering | $378 | $9,034 | — | — | — | $9,412 |
| Customer Service | — | $4,678 | — | $1,156 | — | $5,834 |
| Legal & Compliance | — | — | $3,454 | — | — | $3,454 |
| Business Analytics | — | — | — | — | $2,104 | $2,104 |
| Total | $378 | $13,712 | $3,454 | $1,156 | $2,104 | $20,804 |
Sample data · Department × model report; export CSV by cost center for internal chargeback
Issue a key per department
Each department gets a Virtual Key tied to its cost center, allowed models and monthly budget.
Tag tasks in requests
Add a task tag to each call, such as contract-review — no extra billing code required.
Reports build themselves
Department bills, task costs and cross reports update in real time, ready to export at month end.
IT holds the keys. Teams build with confidence.
Vendor master keys live only in the gateway. Business teams receive controlled Virtual Keys.
- Access and quotas
Issue keys per department, project and environment, each with its own allowed models, budget cap and rate limit. Revoke them anytime when people change roles.
- Automatic retries and fallback models
When an upstream provider rate-limits or times out, the gateway retries according to policy or switches to a preconfigured backup model.
- Call audit trail
Every call records its department, task tag, model, token usage, latency and cost, so reconciliation and tracing are straightforward.
- Request content logging
Turn off logging of prompts and responses to meet your policies, keeping only the data needed for metering.
See where your AI budget is actually going
Book a 30-minute demo. We'll walk through integration and cost reports using your own department structure.
Or email us at admin@contextforgesystem.com