Help & Documentation

What GetWrangler does

GetWrangler sits between your application and your LLM provider. It intercepts every API call, routes it to the least-cost handler, and logs verified savings — without changing your code.

Your proxy endpoint

Point your application at the GetWrangler proxy instead of your LLM provider's API directly.

Your dashboard

Monitor savings, review routing decisions, and export your data at any time.

Integration guide

Step-by-step setup instructions, code examples, and configuration reference.

Evaluation Period

GetWrangler offers a 14-day shadow mode evaluation. During this period, GetWrangler observes your actual traffic, establishes your savings baseline, and shows you exactly what optimization is achievable — before you activate live routing or pay anything.

Option A — Trial Mode (no Abacus account required)

Your account starts immediately using GetWrangler's operator credentials with free model routing (Gemini Flash). This lets you see routing behavior and savings projections right away.

• Limitation — Responses use Gemini Flash regardless of your requested model
• Limitation — Rate limiting may apply on high-volume workloads
• Limitation — Trial is capped at 10,000 requests
• Limitation — Savings baseline reflects free model costs, not your production model

* GetWrangler 1.x provides access to major frontier models (Claude, GPT-4o, Gemini) via a client-owned Abacus.ai account. GetWrangler 2.x will add direct native endpoint support for all major providers.

To upgrade to full evaluation at any time:

1. Log into your dashboard at app.getwrangler.ai, navigate to the Tokens section, and click Create New Token
2. Enter your Abacus API key — it is stored directly in our encrypted vault and never visible to GetWrangler staff
3. Your account upgrades instantly — no operator involvement needed

Option B — Full Evaluation (recommended)

Open a free account at abacus.ai and provide your API key during onboarding OR add it yourself at any time via your dashboard. GetWrangler stores it directly in an encrypted vault.

• Benefit — Real-world response quality matching your production model
• Benefit — No rate limiting beyond your own Abacus plan limits
• Benefit — Accurate savings baseline using your actual model costs
• Benefit — Seamless transition to live routing — no credential changes needed

At the end of your evaluation, visit your dashboard to select a plan and activate live routing. View plan options →

What it is

When your account is in shadow mode, all requests pass through to your LLM provider unchanged. GetWrangler observes your traffic, builds your baseline, and logs what it would have saved — but takes no action.

Why it starts here

Shadow mode lets you verify GetWrangler's behavior before any routing decisions affect your application. Your 14-day evaluation window runs in shadow mode.

When it changes

Add a payment method in your dashboard and self-activate live routing when you are ready. You can see your current mode in the Routing Mode card on your dashboard.

Savings Summary

• Total requests processed this month — count of all API calls routed through GetWrangler
• Total verified savings — actual cost vs. baseline cost, in dollars
• Billing projection — estimated month-end total based on current trajectory

Routing Breakdown

• DETERMINISTIC Matched a rule or template — zero LLM cost
• CACHE Served from cache — zero LLM cost
• LLM Routed to least-cost eligible model
• PASSTHROUGH Passed through unchanged (shadow mode or no cheaper option found)

Usage by API Token

• Cost and request count — broken down by API credential
• Separate tokens per team or service — appear as separate rows for granular attribution
• ⚠ Spike badge — current-day spend exceeds 3× the 7-day trailing average; review for unexpected automation or runaway processes
• Unlabeled tokens — appear as "Unlabeled / Legacy"

Vendor / Model Breakdown

Shows which AI models GetWrangler routed traffic to. Use this to understand your model mix and associated costs across providers.

Routing Mode Status

• Shadow vs. Active — current routing mode indicator
• Source breakdown — routing driven by global default vs. per-request header override

Invoice History

• Monthly billing — platform fee, savings delivered, savings share, total due
• Status — Draft → Reviewed → Sent

Spend Velocity

The SpendVelocity card shows your real-time burn rate at a glance. It updates continuously and projects your trajectory forward so you can act before problems become expensive.

• Current spend rate — rolling daily average compared to your plan period
• Projected monthly total — extrapolated from current velocity
• Burn rate trend — accelerating, steady, or decelerating
• Estimated overage date — the date your projected spend would reach your plan threshold at current velocity

Budget Threshold Alerts

GetWrangler sends email alerts as your spend approaches your plan limit. Alerts are sent once per day to avoid noise.

Anomaly Detection

The ⚠ Spike badge appears on any API token whose current-day spend exceeds 3× the 7-day trailing average. This flags runaway automations and unexpected usage before they compound.

• What triggers it — single-day spend > 3× the trailing 7-day average for that token
• What to do — review recent requests for that token; check for loops, retries, or new automations
• How it clears — automatically on the next daily calculation once the rate normalizes

Contact Management

Map your API tokens to named contacts so every dollar of AI spend is attributed to a specific person, team, or service — not just an anonymous key.

• Token-to-person mapping — label each API token with a name, email, and department
• Department grouping — roll up spend by team or business unit for budget reporting
• Named contacts — alert recipients are tied to the token, not just to the admin

Per-Token Spend Thresholds

Set a daily spend threshold for each API token. When a token's daily spend crosses its threshold, GetWrangler sends an alert email directly to the associated contact.

• Threshold level — set per token in dollars per day
• Alert delivery — email to the named contact for that token
• Daily deduplication — one alert per token per day, regardless of how far over threshold
• No service interruption — thresholds are alert triggers only; requests continue unaffected

Audit & Reporting

Every API call is logged with its token, contact attribution, routing decision, and cost. Export your full spend ledger at any time as CSV for internal reporting or cost allocation.

See Section 7: Exporting Your Data for export details.

Plan tiers

GetWrangler offers four tiers — Starter, Growth, Pro, and Enterprise — sized by the volume of AI spend you want to put under governance. Every tier includes spend velocity monitoring, per-user tracking, department budgets, threshold alerts, anomaly detection, and the savings ledger.

• Starter — $199/mo platform fee · up to 25 users · up to $2,500/mo AI spend governed
• Growth — $499/mo platform fee · up to 100 users · up to $10,000/mo AI spend governed
• Pro — $999/mo platform fee · up to 500 users · up to $50,000/mo AI spend governed
• Enterprise — custom pricing · unlimited users · unlimited spend · dedicated support

ACH vs. Credit Card

GetWrangler charges a savings share — a percentage of verified savings — in addition to the monthly platform fee. Paying by ACH bank transfer gives you a lower savings share rate, meaning you keep more of every dollar saved.

• ACH savings share — 20% (Starter) · 15% (Growth) · 10% (Pro/Enterprise)
• Credit card savings share — 23% (Starter) · 18% (Growth) · 13% (Pro/Enterprise)
• Platform fee — identical for both payment methods

Either way, you keep the substantial majority of every verified dollar saved.

Full plan comparison

See the complete feature-by-feature comparison, FAQ, and trial information.

The Export Data button downloads your complete savings ledger as a CSV file.

• Columns included — timestamp, handler type, baseline cost, actual cost, verified saving, API token
• Row limit — capped at 10,000 rows; the file header indicates if the export is partial
• Use cases — internal reporting, cost allocation, or audit purposes

• Multiple tokens supported — use one token per account, or separate tokens per team, service, or automation pipeline for per-token cost visibility
• Token labels — set via your dashboard; unlabeled tokens appear as "Unlabeled / Legacy"
• Token rotation — contact support@getwrangler.ai to rotate a token; the old token is deactivated immediately upon request
• Per-token thresholds — set daily spend limits and alert contacts per token (see Section 5)

Per-Token Model Preferences

Each token can optionally be set to prefer a specific model for LLM-tier requests.

Three settings are available:

• GetWrangler chooses (default) — maximum cost savings; GetWrangler selects the least-cost eligible model. This is the recommended default for most teams.
• Prefer [model] (soft) — GetWrangler uses your preferred model when the task warrants it, and routes to a lower-cost model when it doesn't. Partial savings preserved.
• Always use [model] (hard) — your preferred model is always used for LLM-tier requests. LLM-tier savings do not apply to this token.

To set a preference: click the arrow (▸) next to any active token in your dashboard Tokens section to expand the model preference selector.

Native Endpoint Support

For teams with fine-tuned models or existing direct endpoint integrations, GetWrangler supports native Anthropic, OpenAI, and Google endpoints directly — so organizations with custom model investments can still benefit from GetWrangler's cost optimization, spend governance, and audit trail without abandoning their existing provider setup. Register up to 5 custom endpoints per account, each with a vault-stored API key and your own contracted pricing rates. Click Save and Test to verify the connection — verified endpoints show a 🔓 icon, unverified or failed ones show 🔒.