ideogram-cost-tuning — Claude Code skill: what it does, how to install
ideogram-cost-tuning controls prepaid Ideogram API consumption with admission ceilings, retry limits, and output tracking. Helps teams eliminate wasted spend on failed or duplicate runs.
MIT 1 installs ★ 2,935 updated 2026-10-29
npx skills add jeremylongshore/tons-of-skills-marketplace --skill ideogram-cost-tuning -a claude-codeWhat ideogram-cost-tuning does
This skill designs budget controls for the Ideogram API without baking in fixed prices that get outdated. It establishes hard admission ceilings per request, tenant, day, or campaign. When limits are hit, intake closes immediately. The workflow prevents duplicate paid work by tracking asynchronous task IDs, expiring stale queue jobs, and trimming retry loops.
Accounting focuses exclusively on safe assets delivered to durable storage. Generations lost to transport drops or flagged by content safety count as budget waste rather than billable milestones. Inspection relies on Read, Glob, and Grep across queues and adapters. Write and Edit handle approved configurations and test cases.
Setup requires a server-side IDEOGRAM_API_KEY, an accountable budget owner, and fresh rate verification from the vendor pricing page. Team keys share a single vendor credit pool, so the skill treats unverified balance assumptions as a failure mode and forbids autonomous balance recharges.
What it helps with
- cap daily and campaign spend across shared Ideogram API keys
- stop duplicate billing on identical image generation requests
- separate billable stored assets from unsafe or abandoned attempts
- block incoming requests automatically before prepaid balance runs dry
Example prompts for the agent
- Audit Ideogram API cost across our batch image generation workers
- Cut Ideogram spend by setting a daily budget ceiling and dropping stale retries
- Budget Ideogram generations for marketing campaigns with a weekly window
How to install ideogram-cost-tuning
1Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill ideogram-cost-tuning -a claude-code2Manually: put the files in place
~/.claude/skills/ideogram-cost-tuning/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/ideogram-cost-tuning/
1Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill ideogram-cost-tuning -a codex2Manually: put the files in place
.agents/skills/ideogram-cost-tuning/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill ideogram-cost-tuning -a cursor2Manually: put the files in place
.cursor/skills/ideogram-cost-tuning/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install jeremylongshore/tons-of-skills-marketplace skills/.curated/ideogram-cost-tuning2Manually: put the files in place
.github/skills/ideogram-cost-tuning/SKILL.md1Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill ideogram-cost-tuning -a gemini-cli2Platform CLI
gemini skills install https://github.com/jeremylongshore/tons-of-skills-marketplace --path skills/.curated/ideogram-cost-tuningManually: put the files in place
.devin/skills/ideogram-cost-tuning/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add jeremylongshore/tons-of-skills-marketplace --skill ideogram-cost-tuning -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about ideogram-cost-tuning
How to install ideogram-cost-tuning in Claude Code?
Run npx skills add jeremylongshore/tons-of-skills-marketplace --skill ideogram-cost-tuning -a claude-code in your terminal.
What does ideogram-cost-tuning do?
It provides budgeting logic and guardrails for Ideogram API usage. The skill configures fail-closed admission rules, deduplicates requests, and counts only durably stored images as successful output.
Can the skill add balance or enable auto-recharge on its own?
No. The skill explicitly forbids topping up credits, activating auto-recharge, or running paid benchmarks without explicit human owner approval.
Source text
SKILL.md · MIT
--- name: ideogram-cost-tuning description: >- Govern Ideogram prepaid spend through admission budgets, route choice, output bounds, useful-result accounting, and cleanup. Use when reducing cost or designing billing controls. Trigger with "cut Ideogram spend", "budget Ideogram generations", or "audit Ideogram API cost". allowed-tools: Read,Glob,Grep,Write,Edit argument-hint: "<workload> <budget-window> <quality-floor>" version: 1.11.0 license: MIT author: Jeremy Longshore <[email protected]> tags: [saas, ideogram, cost] model: inherit effort: high compatibility: "Designed for Claude Code; current prices must be checked at decision time" --- # Ideogram Cost Governance ## Overview Control prepaid Ideogram consumption without embedding prices that can drift. Tie admission to an accountable budget, choose the narrowest route, bound outputs and retries, and measure durable safe results rather than HTTP success or generated candidates. ## Prerequisites - Team billing owner, current balance source, budget period, workload owner, and quality floor. - Request volume, endpoint and rendering mix, output count, retry rate, unsafe rate, and storage failures. - A current pricing review from the first-party API pricing page. ## Current Contract Ideogram API billing is separate from the web application, requires positive prepaid credit, and can use auto-recharge. Keys in one team share credits and billing, so per-key usage is not a separate vendor balance. Prices and recharge choices are time-sensitive and must be verified rather than copied into this skill. ## Authentication Workers use server-side `IDEOGRAM_API_KEY` as `Api-Key` to `https://api.ideogram.ai`. Cost evidence may include team, endpoint, request, and opaque generation identifiers, but not the key, prompts, images, or URLs. ## Instructions 1. Retrieve current first-party pricing and balance behavior, recording date, currency, unit, and owner. 2. Inventory demand by tenant, use case, route, rendering option, output count, retry class, safety result, and durable-storage result. 3. Define request, tenant, daily, campaign, and incident ceilings with fail-closed admission. 4. Choose the least costly route that still meets model, quality, transparency, edit, or tool requirements. 5. Prevent duplicate paid work by persisting async identifiers, bounding retries, expiring stale queue items, and deduplicating callers. 6. Attribute spend to useful safe assets that reached durable storage; surface unsafe, failed, abandoned, and expired-URL waste separately. 7. Canary each tuning change and retain a quality, safety, latency, and rollback comparison. ## Tool Discipline Use Read, Glob, and Grep for billing adapters, queues, metrics, and fixtures. Use Write and Edit for approved budgets, tests, or documentation. Do not add credit, enable auto-recharge, change prices, or run paid experiments by invocation alone. ## Approval Boundaries Require owners for balance additions, recharge settings, budget changes, lower-quality routes, deleted queued work, paid benchmarks, and production rollout. Cost reduction cannot override safety, rights, or retention policy. ## Error Handling - Stop admission before balance exhaustion; repeated failures are not a budget strategy. - Do not retry validation, auth, unsafe, or already accepted async work automatically. - An image that was never durably stored is waste even if generation succeeded. ## Output Return pricing review date, budget model, demand and waste breakdown, selected controls, expected and measured impact, quality and safety guardrails, owner, canary, and rollback. Avoid hard-coded future price claims. ## Examples - Cap a campaign by useful stored assets and expire queued requests when the publication deadline passes. - Report `submitted=100; safe=92; stored=90; duplicates=0; expired_before_submit=14; budget=within-limit`. ## Validation Reconcile application counts with vendor billing evidence, inject budget exhaustion, retry ambiguity, unsafe output, and storage failure, then verify admission closes. Recheck pricing immediately before approving a financial decision. ## Resources - [Current first-party evidence map](references/official-docs.md) — use the dated endpoint, webhook, billing, team, and training links as the contract index for this workflow. - Recheck the endpoint-specific page and current OpenAPI description before relying on an enum, limit, beta feature, or lifecycle claim. - Record live observations as environment-specific evidence, not as universal vendor guarantees.
© Jeremy Longshore · MIT · jeremylongshore/tons-of-skills-marketplace