Atlas uses 1 credit per compiled web page or uncached HTTP retrieval. Adjust the volume sliders to see your projected credit requirements and the ideal tier.
Every boundary, concurrency ceiling, and security feature is explicit and deterministic. Review exact allocations across all tiers.
Feature & Limits
Free
Starter
Pro
Team
Scale
Business
Enterprise
Workload & Compute Limits
Monthly credits allowance
250
5,000
25,000
120,000
480,000
1,000,000
Custom (2B+)
Max concurrent jobs
1
2
5
12
32
48
64+
API rate limit
30 / min
60 / min
120 / min
240 / min
600 / min
900 / min
1,200+ / min
Max pages per crawl
1,000
5,000
25,000
50,000
100,000
200,000
250,000+
Max batch request URLs
100
500
1,000
1,000
1,000
1,000
1,000+
Workspace & Developer Governance
Projects
1
3
10
25
100
200
Unlimited
API keys per workspace
5
10
15
25
50
100
200+
Team members
1
1
5
10
50
100
1,000+
Webhook endpoints
3
5
10
20
50
100
200+
Security & Compliance
API key IP allowlisting
—
—
✓
✓
✓
✓
✓
Audit log export (JSON/CSV)
—
—
—
✓
✓
✓
✓
Single Sign-On (SAML / Okta)
—
—
—
—
—
—
✓ (SAML)
Webhook delivery simulator
✓
✓
✓
✓
✓
✓
✓
Support & Infrastructure
Support tier
Community / Email
Email support
Email support
Priority queue
Priority queue
Priority + Slack
Dedicated Architect
Target uptime SLA
Best effort
99.5%
99.9%
99.9%
99.9%
99.95%
99.99% Custom
Dedicated VPC / Private runners
—
—
—
—
—
—
Optional
Deterministic Billing
Frequently asked questions.
Clear operational guidelines on credits, caching, rate limits, and enterprise deployments.
How are credits counted and deducted?
1 credit corresponds to 1 compiled web document or uncached HTTP retrieval. Fast-path scrape calls (POST /v1/scrape), crawl page extractions, and single-page compilations consume 1 credit per targeted URL. Discovered URLs in site mapping consume 1 credit per 10 URLs mapped.
Do cached requests consume compilation credits?
No. Deterministic web intelligence means reproducible outputs. When an idempotent request replays an identical target within your specified cache TTL, Atlas returns the cached structured IR at 0 credit cost.
What happens if our quota is reached mid-billing cycle?
When allowance is exhausted, API requests receive a structured HTTP 429 response with retry-after headers. You can seamlessly upgrade your tier without service interruption.
What happens when concurrent requests exceed my tier limit?
Atlas queues pending compilation jobs in a persistent queue. Jobs execute immediately as active worker concurrency slots free up; requests are never dropped or errored due to temporary concurrency bursts.
Does the Free tier expire or require a credit card?
No credit card is required. The Free tier does not expire; your 250 compute credits renew automatically every 30-day billing cycle.
Can we switch between plans or cadence mid-cycle?
Yes. Upgrades take effect immediately with prorated billing calculated down to the second via Dodo Payments. Plan changes preserve your API keys, webhooks, and active crawl configurations.
Which payment methods and currencies are supported?
We accept all major credit/debit cards (Visa, Mastercard, American Express), SEPA, UPI, and bank transfers for high-volume Enterprise contracts. All invoices are denominated in USD.
Do you offer dedicated infrastructure or private VPC peering?
Yes. The Atlas Enterprise tier supports private egress IP pools, custom cloud runner deployments (AWS / GCP / Cloudflare), and dedicated data sovereignty controls.
Early Access Opening Soon
Request early access for your team.
We are granting early access in batches to AI engineering teams, search developers, and data infrastructure leads. Secure priority onboarding on your chosen tier.
You're on the early access waitlist!
We'll send your access key and priority onboarding details as soon as the next batch opens.
