Forsig
Policy-first AI gateway
Launch Forsig
Edge-enforced budgets • DB-priced settlement • Full audit trail

Make AI spend predictable. Without slowing teams down.

Forsig is a policy-first AI gateway: issue allowance keys, approve models, reserve preflight, then settle exact usage using DB-based per-model pricing — with an operator-grade audit trail.

Watch the flow
Allowance keys DB pricing Audit Policies
forsig://overview
Live policyEnforced
Allowance keys
Preflight reserve & settle
DB model pricing
Full audit trail
Request story1 API call
POST /v1/chat/completions
Authorization: Bearer FORSIG_ALLOWANCE_KEY

→ Routes to allowed model
→ Enforces policy at edge
→ Reserves budget preflight
→ Settles exact usage
Accurate per-model pricingReserve → Settle flowSpend caps and rate limitsModel allowlistsAudit metricsEdge-friendlyPremium UI dashboardsBudget enforcementAccurate per-model pricingReserve → Settle flowSpend caps and rate limitsModel allowlistsAudit metricsEdge-friendlyPremium UI dashboardsBudget enforcement
What Forsig stands for

Control without friction

Forsig makes budget enforcement invisible to users and obvious to operators — accurate pricing and auditability built in.

Allowance keys

Issue scoped keys that enforce spend rules at the edge.

DB-based pricing

Every model priced from your database — accurate per-token settlement.

Policy enforcement

Reserve preflight, then settle exact usage — no guessing.

Audit clarity

Spend, requests, prompt tokens, completion tokens — visible and exportable.

Model routing

Let users choose approved models while you keep control.

Safer defaults

Design for least-privilege keys and explicit budgets.

Interactive demo

Drag budgets. See the system behave.

A premium product explains itself. Move the sliders and watch how Forsig reserves, settles, and audits.

Interactive budget demo
Drag budget. Watch reserve → settle → audit.
Allowed
Budget$25.00
$2$200
Prompt tokens1,500
Max output tokens800
Preflight reserve
$0.02
prompt + max_output
Settle (actual)
$0.02
output actual: 576
Budget remaining
$24.98
green path
Audit snapshot (animated)
DB pricing
Total spend
$0.02
Requests
1
Prompt tokens
1,500
Completion tokens
576
What’s happening
Forsig enforces policy at the edge.
Allowance keys
Keys carry budget constraints so teams can move fast without surprise spend.
Reserve → settle
Preflight reserve uses prompt + max_output; settlement charges exact usage returned.
DB-based model pricing
Pricing is not hardcoded. Each model’s input/output rate comes from your database.
Audit clarity
Every request is accounted for: spend, tokens, requests, and model usage.
Why it feels “premium”

Motion is used as product feedback: you drag budgets, numbers respond, and the system story becomes obvious without reading docs.

Pricing truth

Model prices that feel alive

Switch models and watch estimates adapt — mirroring DB-first pricing in the real gateway.

Model pricing (animated)
Switch models — costs reflow instantly.
Input price
$5.00 / 1M
prompt tokens
Output price
$15.00 / 1M
completion tokens
Example workloads
cost preview
WorkloadPromptCompletionEst. cost
Small chat800300$0.01
Agent step2,5001,100$0.03
Tool-heavy6,0002,500$0.07
In production, Forsig reads these rates from your DB and settles using actual usage returned.
Why DB pricing matters
Accuracy is a trust feature.
One source of truth
Update a model’s rate once — all calculations follow.
Fewer “gotchas”
No drifting env values. No mismatched dashboards.
Auditable to the cent
Reserve/settle values match what you store & report.
Premium UX rule

The UI should show what operators care about: policy outcomes, pricing truth, and audit coverage.

Product trailer

Scroll through the Forsig flow

A scroll-driven storyline: keys → preflight reserve → settle from DB → audit trail.

Product trailer
scroll to play
Keys → reserve → settle from DB → audit trail
Issue allowance keys
Give teams scoped keys with budgets. Let them build fast without surprise spend.
Reserve preflight
Reserve prompt + max_output before execution. If budget fails, block early.
Settle exact usage
After usage returns, settle precise cost using DB-first model pricing.
Audit everything
Every request is logged: spend, tokens, metadata, outcomes — operator grade.
Issue allowance keys
Give teams scoped keys with budgets. Let them build fast without surprise spend.
Least-privilege access
Per-key budget caps
Model allowlists
System view
enforced
Policy
Key budgets + allowlist
Outcome
Scoped access issued