AIVAX
Console

Plans & pricing

The right plan for the stage you are in.

Start with daily service allowances at no cost, even on Free. Build with included RAG embeddings, Reflex reranking, Julia-1 decisions and OCR extraction, then unlock more capacity as you grow. Model inference and service usage outside these allowances are billed separately.

Find your starting point

Choose your stage

A path from first request to production scale.

Choose by the limit you are likely to reach first, not just the monthly fee. These subscription prices are a planning snapshot; check current terms in the console before upgrading.

01 / Explore

Free

Build your first working prototype with free daily service allowances and no monthly subscription.

$0/mo No monthly subscription Start on Free

At a glance

RAG search embeddings
≈ 400 short queries/day included†
Reflex reranking
Daily reranking allowance included
Julia-1 decisions
Daily decision allowance included
OCR / Fetch extraction
Daily extraction allowance included
Integrated model requests
20/min · 500/day
Models
Basic / low-price
Storage
30 MB · no expansion

Move to Pro when daily request limits or the 30 MB storage ceiling start shaping the product.

02 / Launch

Pro

For shipping a live application, testing more capable models and serving a growing user base.

$19/mo Subscription + metered usage Choose Pro

At a glance

RAG embeddings
15× the Free allowance
Reflex reranking
4× the Free allowance
Julia-1 decisions
2× the Free allowance
OCR / Fetch extraction
8× the Free allowance
Integrated model requests
150/min
Models
Full catalog
Storage
2 GB included · $0.50/GB/month excess

Move to Max when throughput, storage and retention become operating requirements rather than occasional spikes.

03 / Scale

Max

For teams running production volume that need more throughput, more storage and a longer audit trail.

$99/mo Subscription + metered usage Choose Max

At a glance

RAG embeddings
50× the Free allowance
Reflex reranking
20× the Free allowance
Julia-1 decisions
3× the Free allowance
OCR / Fetch extraction
25× the Free allowance
Integrated model requests
1,000/min
Models
Full catalog · 2% inference uplift
Storage
10 GB included · $0.30/GB/month excess

Move to Ultimate when fixed throughput ceilings, retention or support level start limiting the operation.

04 / Ultimate

Ultimate

For high-volume operations that need the largest allowances and dedicated support.

$199/mo Subscription + metered usage Choose Ultimate

At a glance

RAG embeddings
100× the Free allowance
Reflex reranking
50× the Free allowance
Julia-1 decisions
5× the Free allowance
OCR / Fetch extraction
50× the Free allowance
Integrated model requests
No plan cap*
Models
Full catalog · no inference uplift
Storage
20 GB included · $0.20/GB/month excess

Choose Ultimate for the largest daily allowances, 30-day retention and dedicated support.

05 / Enterprise

Enterprise

For teams whose volume does not fit a fixed tier. Limits and terms are sized to your workload and agreed with the AIVAX team.

Custom Negotiated subscription + metered usage Talk to us

At a glance

Rate limits
Custom limits scaled to your expected throughput
Models
Full catalog
Pricing
Negotiated monthly price

Talk to us when Ultimate limits become the constraint, or when you need terms agreed in advance.

† Illustrative daily estimates, not measured averages or guaranteed request limits. The RAG scenario assumes 20 input tokens per query. Reflex and Julia-1 allowances are shown without request estimates pending representative workload measurements. Estimates are rounded down with headroom and do not count the extra usage margin. RAG estimates assume no document insertion that day and cover query embeddings only, not reranking or answer generation. Actual capacity varies with input size and content; balance requirements and rate limits still apply. OCR capacity depends on processing time, so no page count is promised.

* Model, provider, gateway and endpoint limits may still apply. Request ceilings vary by model rate-limit group. BYOK requests have separate limits.

What changes as you grow

Daily allowances. Included even on Free.

Every Free account gets daily allowances for RAG embeddings, Reflex, Julia-1 and OCR at no extra cost. Pro, Max and Ultimate give you more included capacity as your workload grows. Compare your room to build below.

CapabilityFreeProMaxUltimate
OCR / Fetch extraction includedFree daily allowance8× the Free allowance25× the Free allowance50× the Free allowance
RAG search and insertion embeddings included≈ 400 short queries/day†15× the Free allowance50× the Free allowance100× the Free allowance
Reranking (Reflex) includedFree daily allowance4× the Free allowance20× the Free allowance50× the Free allowance
Semantic decisions (Julia-1) includedFree daily allowance2× the Free allowance3× the Free allowance5× the Free allowance
Integrated model requests20/min · 500/day150/min1,000/minNo plan cap*
Inference uplift25%5%2%None
Semantic searches20/min300/min1,500/min3,000/min
Batch workflow items500/day25,000/day250,000/dayNo plan cap*
Web searches15/day500/day4,000/day10,000/day
Storage included30 MB2 GB10 GB20 GB
Conversation retention2 hours2 days14 days30 days
SupportEmailPriorityPriorityDedicated

† See the estimate assumptions; multipliers compare allowances, not request counts. Uncovered OCR / Fetch extraction costs $0.15, $0.05, $0.03 or $0.02 per 1,000 PUs on Free, Pro, Max or Ultimate respectively. Optional JSON conversion is billed separately.

Free, Pro, Max and Ultimate each include separate daily allowances for extraction, RAG search and insertion embeddings, reranking and semantic decisions. These allowances are not transferable between services, and unused amounts do not carry over. Custom has no included allowances. RAG response tokens and model inference (LLM) are metered normally and are not covered by these allowances. Reranking coverage currently applies to Reflex and decisions coverage to Julia-1.

Coverage applies per metered service item (document embedding, query-term embedding, reranking call, decision input, extraction): each item is covered in full if consuming it keeps daily usage within the service's allowance plus a 10% margin; otherwise that item is fully billed and leaves the allowance untouched, so one request can mix covered and billed items. Balances and rate limits still apply.

See all limits

Understand the bill

The subscription is not the whole invoice.

A practical estimate starts with the plan, then adds the services your application actually uses.

01 / Models

Choose the model and measure tokens.

Input, output and media rates vary by model. The plan's inference uplift applies to integrated-model usage; check live rates in the model catalog.

02 / Data

Account for RAG and storage.

Indexing, search, reranking and answer generation are distinct operations. Paid-plan storage above the included quota is billed by usage over time.

03 / Your keys

BYOK keeps the model bill with your provider.

AIVAX request limits and metered platform services still apply. Paid, unexpired account credit covers billable AIVAX usage.

Review service rates and billing terms

Start small. Upgrade when the workload makes the case.

Open the console