Pricing

from $3.00 / 1,000 ultra api call (validated + fallback protected)s

OpenRouter Ultra – Fix Broken JSON & Structured Output

Stop dealing with broken JSON from LLMs. OpenRouter Ultra delivers reliable structured outputs from text or datasets using automatic retries, fallback models, and built-in repair. Get predictable costs and production-ready results for batch processing at scale.

Pricing

from $3.00 / 1,000 ultra api call (validated + fallback protected)s

Rating

0.0

(0)

Developer

Chris

Actor stats

Bookmarked

Total users

Monthly active users

17 days ago

Last modified

Categories

Developer tools

Automation

Execution Mode

mode

Required

Choose 'api' to run a Standby server, or 'batch' to process a dataset.

Type:string

Default:api

Options:

apibatch

OpenRouter API Key (BYOK)

openRouterApiKey

Optional

Your OpenRouter API Key. This is stored securely.

Type:string

Model Fallback Chain

modelChain

Optional

If the first model fails, the Actor will automatically retry with the next in the list.

Type:string[]

Default:

[
  "openai/gpt-4o-mini",
  "anthropic/claude-3-haiku"
]

Policy Preset

policyPreset

Optional

Apply a predefined configuration profile. ultra-fast minimizes latency, ultra-smart enables repair + debug, ultra-private enforces PII masking and zero data retention, ultra-json enables strict JSON mode.

Type:string

Default:custom

Options:

ultra-fastultra-smartultra-privateultra-jsonultra-batchultra-extractultra-debugcustom

Task Class

taskClass

Optional

Hint for the smart routing engine. 'auto' infers the task from the request shape.

Type:string

Default:auto

Options:

autochatextractionclassificationreasoningembedding

Target API Route

route

Optional

Which OpenRouter API route to target.

Type:string

Default:chat

Options:

chatresponsesanthropicembeddings

Max Cost Per Run (USD)

maxCostPerRunUsd

Optional

Hard budget limit. The Actor will stop processing if this cost is exceeded.

Type:number

Minimum:0

Default:5

Max Cost Per Request (USD)

maxCostPerRequestUsd

Optional

Maximum allowed cost for a single API request.

Type:number

Minimum:0

Timeout (ms)

timeoutMs

Optional

Per-model timeout in milliseconds.

Type:integer

Minimum:1000

Default:30000

Latency Target (ms)

latencyTargetMs

Optional

Target latency hint for the routing engine. Lower values prefer faster models.

Type:integer

Minimum:100

Max Retries

maxRetries

Optional

Maximum number of retries per request.

Type:integer

Minimum:0

Default:2

Enable Exact Response Cache

enableExactCache

Optional

If enabled, identical requests will be served from the internal cache (SHA-256 hash), saving tokens and money.

Type:boolean

Default:true

Cache TTL (seconds)

cacheTtlSeconds

Optional

How long cached responses remain valid.

Type:integer

Minimum:1

Default:86400

Cache Namespace

cacheNamespace

Optional

Isolate cache entries by namespace. Useful for multi-tenant setups.

Type:string

Default:default

Enable Prefix Cache

enablePrefixCache

Optional

Enable advanced prefix-based caching for long prompts.

Type:boolean

Default:true

User Scoped Cache

userScopedCache

Optional

Isolate cache by user ID (if provided in request headers).

Type:boolean

Default:true

Enable JSON Self-Healing

jsonStrictMode

Optional

Automatically strip markdown and repair broken JSON syntax from the AI's output.

Type:boolean

Default:true

JSON Schema

jsonSchema

Optional

Validate LLM output against this JSON Schema. Used with JSON strict mode.

Type:object

Enable Deterministic Repair

enableDeterministicRepair

Optional

Try regex-based JSON repair before falling back to model-based repair.

Type:boolean

Default:true

Max Repair Attempts

maxRepairAttempts

Optional

How many times to attempt repair before failing.

Type:integer

Minimum:0

Default:1

Enable Repair Model

enableRepairModel

Optional

Use a specialized model to fix JSON that deterministic repair cannot handle.

Type:boolean

Default:false

Repair Model Name

repairModel

Optional

Model to use for JSON self-healing if deterministic repair fails.

Type:string

Default:openai/gpt-4o-mini

Enable Debug Artifacts

enableDebugArtifacts

Optional

Store JSON parse/validation failure details for post-run inspection.

Type:boolean

Default:true

Enable PII Masking

piiMasking

Optional

Scrub emails and phone numbers before sending data to OpenRouter.

Type:boolean

Default:false

Zero Data Retention Only

zdrOnly

Optional

Only route to providers that support zero data retention policies.

Type:boolean

Default:false

Redact Logs

redactLogs

Optional

Redact sensitive data from Actor logs.

Type:boolean

Default:true

Target Dataset ID (Batch Mode Only)

datasetId

Optional

The Apify Dataset ID containing the rows you want to process.

Type:string

Request Template (Batch Mode Only)

requestTemplate

Optional

The OpenRouter payload template. Use {{fieldName}} to inject data from your dataset.

Type:object

Input Field Mapping

inputMapping

Optional

Advanced mapping of dataset fields to template variables.

Type:object

Batch Sample Size

sampleSize

Optional

Process only the first N rows (useful for testing).

Type:integer

Minimum:1

Default:10

Dead Letter Dataset ID

deadLetterDatasetId

Optional

Failed batch rows are pushed to this dataset for inspection.

Type:string

Completion Webhook URL

webhookUrl

Optional

URL to POST a summary to when batch processing completes.

Type:string

OpenRouter Models Scraper

automation-lab/openrouter-models-scraper

Scrapes all AI models from the OpenRouter platform including pricing, context length, capabilities, and supported parameters.

Stas Persiianenko

OpenRouter - Bring Your Own Key

superlativetech/open-router

Use the OpenRouter platform to choose the best and most cost effective model for your prompts utilizing a standardized interface (including ChatGPT, Claude, Gemini, Llama, Mistral, and more). See instructions for creating an OpenRouter account and API key.

Superlative

Broken Link Checker

jancurn/find-broken-links

Crawls a website and finds broken links. Unlike other similar SEO analysis tools, the actor also reports broken URL #fragments. The results are stored in a JSON and HTML report.

Jan Čurn

779

Broken Link Finder

pillowy_travel/broken-link-finder

Finds and analyzes broken links on given web pages

SAHIL KUMAR

Shopify Scraper - Ultra Fast Product Extractor

kalirobot/shopify-scraper

Extract full product data from any Shopify store in seconds: titles, descriptions, prices, variants, stock, images and more. Ultra fast, reliable and API ready.

Dennis

5.0

AI Selector Repair Tool

automation-lab/ai-selector-repair

Fix broken CSS and XPath selectors with AI. Provide a URL and broken selectors — get back updated, working selectors with confidence scores and explanations of what changed in the page structure.

Stas Persiianenko

Broken Link Checker - Ensure Your Website's Integrity

dainty_screw/find-broken-links-of-your-website

Maintain your website's health and user experience with our Broken Link Checker. Easily identify and fix broken links to enhance your site's navigation, improve SEO, and keep visitors engaged.

codemaster devops

5.0

Broken Link Checker

automation-lab/broken-link-checker

Broken Link Checker crawls your website, discovers all internal and external links, and verifies each one. It finds 404 errors, server errors, timeouts, and other broken links — then tells you exactly which page links to each broken URL and what the anchor text says.

Stas Persiianenko

⚡ Ultra Fast Linkedin Ad Library Scraper ⚡

operatorultra/Linkedin-Ad-Library-Scraper

Extract competitor intelligence and campaign insights at scale in no-time.

Operator Ultra

Agent Output Cleaner

yauponlabs/structured-output-cleaner

Fix messy LLM and agent outputs into clean, machine-usable JSON. Repair common structured-output breakage, return a repair report, and fail honestly on ambiguous cases.