static audit · 2026-10-26T04:02:06.098Z

mendableai/firecrawl

120 source files scanned · 38 LLM call site(s) · ~13% of LLM spend looks recoverable (static heuristics)

$13.5per 10k runs today (gemini-2.5-flash, ~2000 prompt tokens)
$11.75per 10k runs with fixes applied

badge   ![TokenZero audit](https://tokenzero.sh/v1/badge/audit/aud_4mVBLWVqlxxBh6DLl3bx3Q.svg)

repo badge   [![LLM spend audited by TokenZero](https://tokenzero.sh/v1/badge/repo/mendableai/firecrawl.svg)](https://tokenzero.sh/audit/mendableai/firecrawl)

Models referenced

gemini-2.0-flash ×6
gemini-2.5-flash-image-preview ×5
gemini-2.5-pro-exp-03-25 ×5
claude-3-7-sonnet-20250219 ×4
gpt-4.1-mini ×3
gpt-5 ×2 (cheaper sibling: gpt-5-mini)
claude-3-5-sonnet-20240620 ×1
gpt-4o ×1 (cheaper sibling: gpt-4o-mini)
gpt-4o-mini ×1

Findings

SeverityWhereFindingEst.
highapps/api/src/scraper/scrapeURL/transformers/llmExtract.ts:860gpt-5 used for a task whose prompt reads like extraction/classification
Full-size pricing ($1.25/M in, $10/M out) on a constrained task. gpt-5-mini costs ~80% less per call at this shape.
Fix: Opt into TokenZero cascade routing (`X-TokenZero-Cascade: auto`) so this call is routed to gpt-5-mini only when the classifier is confident (an opt-in shadow parity monitor is coming; nothing is promised), or switch the model here behind an eval.		80%
mediumexamples/gemini-2.5-crawler/gemini-2.5-crawler.py:195gemini-2.5-pro used for a task whose prompt reads like extraction/classification
Frontier pricing ($1.25/M in, $10/M out) on a constrained task. gemini-2.5-flash costs ~75% less per call at this shape.
Fix: Opt into TokenZero cascade routing (`X-TokenZero-Cascade: auto`) so this call is routed to gemini-2.5-flash only when the classifier is confident (an opt-in shadow parity monitor is coming; nothing is promised), or switch the model here behind an eval.		75%
mediumexamples/gemini-2.5-crawler/gemini-2.5-crawler.py:313gemini-2.5-pro used for a task whose prompt reads like extraction/classification
Frontier pricing ($1.25/M in, $10/M out) on a constrained task. gemini-2.5-flash costs ~75% less per call at this shape.
Fix: Opt into TokenZero cascade routing (`X-TokenZero-Cascade: auto`) so this call is routed to gemini-2.5-flash only when the classifier is confident (an opt-in shadow parity monitor is coming; nothing is promised), or switch the model here behind an eval.		75%
mediumexamples/gemini-2.5-web-extractor/gemini-2.5-web-extractor.py:59gemini-2.5-pro used for a task whose prompt reads like extraction/classification
Frontier pricing ($1.25/M in, $10/M out) on a constrained task. gemini-2.5-flash costs ~75% less per call at this shape.
Fix: Opt into TokenZero cascade routing (`X-TokenZero-Cascade: auto`) so this call is routed to gemini-2.5-flash only when the classifier is confident (an opt-in shadow parity monitor is coming; nothing is promised), or switch the model here behind an eval.		75%
mediumapps/api/src/lib/deterministicJson/llm/prompts.ts:384 static prompt template(s), ~2,843 tokens, and no caching directive anywhere in the repo
Most call sites use providers with automatic prefix caching, which only pays off when these templates are sent as a byte-identical prefix: static template first, nothing volatile (dates, ids, user data) interpolated into it.
Fix: Keep templates first and static; TokenZero adds a stable `prompt_cache_key` so related requests land on the same cache.		12%
mediumapps/api/src/scraper/scrapeURL/transformers/llmExtract.ts:508"gpt-4.1-mini" retires on 2027-01-14
The provider announced on 2026-10-14 that this model retires on 2027-01-14 (80 days). Recommended replacement: gpt-6-luna.
Fix: Move to gpt-6-luna (eval first); TokenZero's cascade can route to it, and the gateway warns in X-TokenZero-Warning until you do.		0%
mediumapps/api/src/scraper/scrapeURL/transformers/llmExtract.ts:860"gpt-5" retires on 2026-12-11
The provider announced on 2026-06-11 that this model retires on 2026-12-11 (46 days). Recommended replacement: gpt-5.6-sol.
Fix: Move to gpt-5.6-sol (eval first); TokenZero's cascade can route to it, and the gateway warns in X-TokenZero-Warning until you do.		0%
lowexamples/gemini-2.5-screenshot-editor/cli.py:287LLM call inside a loop
Per-item calls repeat the full prompt prefix for every element. If items are independent and latency-tolerant, the provider Batch API is 50% cheaper, and OpenAI Flex processing (`service_tier: "flex"`) gives the same 50% on ordinary calls without restructuring; if the prefix is shared, caching pays off immediately.
Fix: Batch where possible, or set `service_tier: "flex"` on OpenAI calls that can wait; otherwise keep the shared prefix cache-formatted (TokenZero does this automatically).		25%
lowexamples/claude-3.7-stock-analyzer/claude-3.7-stock-analyzer.py:127LLM call inside a loop
Per-item calls repeat the full prompt prefix for every element. If items are independent and latency-tolerant, the provider Batch API is 50% cheaper, and OpenAI Flex processing (`service_tier: "flex"`) gives the same 50% on ordinary calls without restructuring; if the prefix is shared, caching pays off immediately.
Fix: Batch where possible, or set `service_tier: "flex"` on OpenAI calls that can wait; otherwise keep the shared prefix cache-formatted (TokenZero does this automatically).		25%
lowexamples/claude_stock_analyzer/claude_stock_analyzer.py:127LLM call inside a loop
Per-item calls repeat the full prompt prefix for every element. If items are independent and latency-tolerant, the provider Batch API is 50% cheaper, and OpenAI Flex processing (`service_tier: "flex"`) gives the same 50% on ordinary calls without restructuring; if the prefix is shared, caching pays off immediately.
Fix: Batch where possible, or set `service_tier: "flex"` on OpenAI calls that can wait; otherwise keep the shared prefix cache-formatted (TokenZero does this automatically).		25%
lowexamples/openai_swarm_firecrawl_web_extractor/main.py:61"gpt-4o" retires on 2027-01-14
The provider announced on 2026-10-14 that this model retires on 2027-01-14 (80 days). Recommended replacement: gpt-6-luna.
Fix: Move to gpt-6-luna (eval first); TokenZero's cascade can route to it, and the gateway warns in X-TokenZero-Warning until you do.		0%
lowexamples/openai_swarm_firecrawl/main.py:24"gpt-4o-mini" retires on 2027-01-14
The provider announced on 2026-10-14 that this model retires on 2027-01-14 (80 days). Recommended replacement: gpt-6-luna.
Fix: Move to gpt-6-luna (eval first); TokenZero's cascade can route to it, and the gateway warns in X-TokenZero-Warning until you do.		0%
infoexamples/gemini-2.0-crawler/gemini-2.0-crawler.py:58Unrecognised model id "gemini-2.0-flash"
Not in the current pricing table: it may be retired or renamed. Retired models return errors or are silently remapped by some providers.
Fix: Check the provider model list; TokenZero flags unpriced models in `X-TokenZero-Warning`.		0%
infoexamples/claude3.7-web-crawler/claude3.7-web-crawler.py:39Unrecognised model id "claude-3-7-sonnet-20250219"
Not in the current pricing table: it may be retired or renamed. Retired models return errors or are silently remapped by some providers.
Fix: Check the provider model list; TokenZero flags unpriced models in `X-TokenZero-Warning`.		0%
infoexamples/claude_stock_analyzer/claude_stock_analyzer.py:127Unrecognised model id "claude-3-5-sonnet-20240620"
Not in the current pricing table: it may be retired or renamed. Retired models return errors or are silently remapped by some providers.
Fix: Check the provider model list; TokenZero flags unpriced models in `X-TokenZero-Warning`.		0%

Fix all of it with one line

OPENAI_BASE_URL=https://tokenzero.sh/v1
ANTHROPIC_BASE_URL=https://tokenzero.sh

Keys pass through untouched. Cache formatting and exact caching are on by default; cascade routing stays opt-in. How it works