Start free. Scale securely.

Get started for free or upgrade for premium features.

Start free. Scale securely.

Get started for free or upgrade for premium features.

Free

Developer

For small teams to get started developing their first AI features.

Create account

Included

1 user

50k spans/month

1GB processed data

10 MB memory storage

3 agents

50 agent runs/month

3 deployments

1 webhook

2 knowledge bases/memory stores

14 day retention

50 API calls/day

Core features

AI gateway

Agent runtime

Prompt engineering

Online & offline evals

Experimentation

Observability

Deployments

Knowledge bases

Platform APIs

Analytics & insights

Support

Support center

Compare our plans

Free

Developer

For small teams to get started developing their first AI features.

Create account

Included

1 user

50k spans/month

1GB processed data

10 MB memory storage

3 agents

50 agent runs/month

3 deployments

1 webhook

2 knowledge bases/memory stores

14 day retention

50 API calls/day

Core features

AI gateway

Agent runtime

Prompt engineering

Online & offline evals

Experimentation

Observability

Deployments

Knowledge bases

Platform APIs

Analytics & insights

Support

Support center

Compare our plans

Pay as you go

Growth

For teams ready to scale AI features in production.

Create account

Everything in Developer, plus:

Included

Unlimited users

€35 per seat per month

100k spans/month

thereafter €7 / 100k spans

1 GB processed data

thereafter €3 / GB

10 MB Memory storage

thereafter €0.10 / MB

Unlimited agents

500 agent runs/month

thereafter €0.01 per run

Unlimited deployments

5 webhooks

2 knowledge bases/memory stores

buy add on for additional usage

30 day retention

Higher rate limits

Core features

Unlock all feature usage

Share/clone agent templates

PII filtering

Support

Email support

Priority SLA

KB / Memory Stores Add-on

€500/mo

Unlimited Retrievals

Ingestion

Parsing

Chunking

2.5 GB Document processing

therafter €0.10 per MB

optional

Teams Add-on

€300/mo

Enterprise SSO (e.g. Okta, Microsoft)

Authentication: SAML/OIDC

SSO enforcement

Fine-grained RBAC

Support via dedicated Slack channel

optional

Compare our plans

Custom

Enterprise

For large organizations with custom requirements and heavy AI workloads.

Book a demo

Everything in Growth, plus:

Included

Custom seats

Custom spans

Custom processed data

Custom memory storage

Unlimited agents

Unlimited agent runs

Unlimited deployments

Unlimited webhooks

Custom knowledge bases

Custom retention period

Custom rate limits

Premium features

Enterprise API

Role-based access control

Document processing priority

SSO / SCIM API

Audit logs

SOC 2 report

HIPAA BAA and Custom DPA

Premium support

Enterprise SLA

Slack / Teams support

Dedicated Solutions Engineer

Deployment options

AWS / Azure Marketplace

On-prem or private cloud

Compare our plans

Custom

Enterprise

For large organizations with custom requirements and heavy AI workloads.

Book a demo

Everything in Growth, plus:

Included

Custom seats

Custom spans

Custom processed data

Custom memory storage

Unlimited agents

Unlimited agent runs

Unlimited deployments

Unlimited webhooks

Custom knowledge bases

Custom retention period

Custom rate limits

Premium features

Enterprise API

Role-based access control

Document processing priority

SSO / SCIM API

Audit logs

SOC 2 report

HIPAA BAA and Custom DPA

Premium support

Enterprise SLA

Slack / Teams support

Dedicated Solutions Engineer

Deployment options

AWS / Azure Marketplace

On-prem or private cloud

Compare our plans

Trusted by

hear.com
hear.com
hear.com
hear.com

Compare our plans

Developer

Get started

Growth

Get started

Enterprise

Book a demo

Usage & Storage

Users

1

Unlimited

€35/seat per month

Custom

Spans

50k/month

100k/month

thereafter €7/100k spans

Custom

Agents

3

Unlimited

Unlimited

Agent Runs

50 runs/month

500 runs/month

Custom

Knowlegde Bases/Memory Stores

2

2

Custom

KB/Memory Store storage

10 MB

10 MB

Custom

Trace retention

14 days

30 days

Custom

Ingestion volume

1 GB

1 GB

Custom

Rate limits

50/day

Higher

Custom

Platform API

Enterprise API

AI Gateway

Multi-modality

Multi-modality

Multi-modality

Agents

Build single and multi-agent systems

Exposable via A2A protocol

Add tools with remote MCP server

Create and add HTTP, Function, Python, and JSON Tools

Trigger agent via API

Use preferred model

Share and clone agent templates

Unlimited agents

Prompt

Prompt library

Prompt engineering

Structured outputs

Playgrounds

Multi-prompt comparison

Evaluator library

Deploy

Deployments

3

Unlimited

Unlimited

Retries & fallbacks

Versioning

Contextual rules engine

LLM cache

File ingestion

Experimentation

Agent experimentation

Prompt experimentation

Experimentation via API

CI/CD

Knowledge Bases & Memory Stores

Storage

10MB

10 MB

Custom

RAG-as-a-Service

Chunk explorer

Embedding & reranking

RAG evaluators

Document processing priority

Observe

Real-time traces

Threads tracking (conversation, sessions, users)

Cost and token tracking

OpenTelemetry

Trace retention

14 days

30 days

Custom

Dashboards

Online evaluators

Guardrails

Experiment exports

Webhooks

1

5

Unlimited

Optimize

Human evaluations

Corrections

Feedback API

Dataset curation

Security & privacy

PII filtering

SOC 2 report

Role-based access control

AWS/Azure marketplace

VPC deployment

SSO / SCIM API

HIPAA

Audit Logs

Custom rate limits

High rate limits

Custom

Uptime SLA

Support

Support center

Email support

Slack/Teams support

only with Teams add-on

Dedicated account manager

Solutions engineer

SLA

Future-proof solution

Why teams switch

One control tower across teams

Unite engineering, product, and data teams in one place. Shared truth, role-based workflows, and human-in-the-loop feedback that drives continuous improvement.

Sign up now

Deploy anywhere, safely

Our cloud, your cloud, or your servers. Private connections supported. Roll out safely and roll back fast.

See deployment options

Compliant, secure and flexible

SOC 2-certified, GDPR-compliant, and aligned with the EU AI Act. Manage risk responsibly with EU or US data residency and regional storage and processing across open and closed ecosystems.

FAQ

Frequently asked questions

What is the easiest way to try Orq.ai?

You can sign up for a free Developer account to start building with your own data. The free plan includes 50k spans per month, and access to all core platform features with some restrictions — no credit card required. See the comparison table above for full details.

What is a span?

A span is the fundamental unit of observability in Orq.ai. Each span represents a discrete operation in your AI application, such as an model invocation, a tool execution, a retrieval step, or an evaluation. Spans are the building blocks of traces and are automatically created when your application interacts with Orq.ai.

What counts as a span?

A span is recorded whenever a discrete operation occurs within your AI application through Orq.ai. This includes:

  • Each request routed through Orq.ai's AI Gateway (LLM calls, generative model invocations)

  • Individual steps within an agent's execution (tool calls, reasoning steps, retrieval)

  • External spans you send from your own infrastructure via the SDK or API

  • Evaluator and guardrail executions

  • Feedback and scoring events

What is the difference between a span and a trace?

A trace represents a complete end-to-end interaction in your application, such as a user request or an agent invocation. A span is a single step within that trace. For example, when an agent processes a user query, the full interaction is captured as one trace, while each individual operation within it (an LLM call, a tool execution, a retrieval step) is recorded as a separate span. Billing is based on spans.

What is an agent run?

An agent run is recorded each time an agent is invoked through Orq.ai's Agent Runtime. A single agent run may contain multiple spans (LLM calls, tool executions, memory lookups, etc.), giving you full visibility into every step of the agent's behavior.

What does processed data mean?

Processed data is the total volume of data ingested by Orq.ai, including inputs, outputs, prompts, and metadata. It is measured on disk and resets each billing cycle. For example, 100k requests with a 50-page PDF attached will generate far more processed data than 100k single-sentence requests — volume depends on the size of your data, not just the number of requests.

What is a knowledge base, what is a memory store, and what does "2 free" mean?

Knowledge bases and memory stores are both built on top of a vector database but serve different purposes. A knowledge base is fully controlled by the builder — you upload and manage documents, and can expose them on deployments and agents for retrieval-augmented generation (RAG). A memory store is filled dynamically by agents at runtime, such as conversation history or contextual memory. The free plan includes 2 combined — that could be 2 knowledge bases, 2 memory stores, or 1 of each.

What is an agent, and how are agents counted?

An agent in Orq.ai is an AI system you build and run through the Agent Runtime, with support for tools, memory, multi-step reasoning, and orchestration. Agents are counted separately from deployments. The number of agents you can create depends on your plan.

What is a deployment?

A deployment exposes a prompt and model configuration to the Orq.ai API so your application can call it in production. Deployments include versioning, retries, fallbacks, and contextual rules. The free plan includes 3 deployments.

What does a higher rate limit mean in the paid plan?

The paid plan offers higher API rate limits, allowing your applications to handle more concurrent requests. This is important for production workloads with higher traffic. Exact limits depend on your plan.

What are platform APIs and Enterprise APIs?

The platform API is the standard Orq.ai API used to invoke deployments, agents, manage prompts, log traces, and interact with the platform programmatically. The Enterprise API extends this with additional capabilities such as SCIM provisioning, audit logs, and bulk operations for larger organizations with advanced administration needs.

What is a playground?

A playground is a no-code workspace for rapidly iterating on prompts and model configurations. You can compare model and prompt behavior side-by-side, fine-tune parameters, and share configurations with teammates — all before moving to a deployment or experiment.

What does document processing priority mean?

Document processing priority determines how quickly your files are parsed and indexed when uploading documents to a knowledge base. Enterprise customers receive priority processing, meaning their documents are chunked and made searchable faster.

Which plan is right for me?

  • Developer Free: For individual developers or single teams getting started. Includes 50k spans/month and access to core platform features with some restrictions. See the comparison table above for details.

  • Developer Paid: For small teams that need higher limits. Includes 100k spans/month, with additional spans at €7 per 100k. Unlocks higher usage limits and longer retention.

  • Enterprise: For larger organizations with custom requirements such as high data volumes, advanced security needs, on-premise deployment, dedicated support, and role-based access control.

When do I get billed?

You are billed at the end of your monthly billing cycle. All included allowances reset at the start of each new cycle.

What happens if I exceed my included limits?

Any usage beyond your included allowances is billed against metered pricing and appears as separate line items on your invoice. You can review the metered rates for spans, agent runs, and processed data in the pricing table above.

Can I manage or cancel my subscription?

Yes, you can manage or cancel your paid Developer subscription at any time. Changes take effect at the end of your current billing cycle. For Enterprise plans, please contact our team.

Can I use my own private or fine-tuned models?

Yes! All Orq.ai plans support private and fine-tuned models. You can connect your own models through the AI Router alongside 300+ supported models from 20+ providers. For more information, feel free to contact our team.

Does Orq.ai support self-hosted or on-premise deployment?

Yes, self-hosted and on-premise deployment options are available for Enterprise customers. This includes deployment in your own VPC or private cloud to meet stricter compliance and data residency requirements. Contact us to discuss your setup.

What security and compliance standards does Orq.ai meet?

Orq.ai is SOC 2 Type II certified, GDPR compliant, and aligned with the EU AI Act, ensuring enterprise-grade security and data protection. For organizations with advanced security needs, we offer self-hosted, VPC, or hybrid deployment options. Visit our Trust Center at trust.orq.ai for full details.

Can I redline contracts?

Yes, we offer customized contracts for Orq.ai Enterprise customers with a yearly commitment. Please contact us for more details. The Developer plans are designed to be self-serve on our standard terms.

I have security questions, where do I start?

We publish our security documentation and controls publicly. Please visit our Trust Center at trust.orq.ai for full details, or contact our team directly.

How does seat-based billing work when adding or removing team members?

Seats are automatically adjusted: added when you invite a member, and removed (after the current invoice is paid) when you remove one. When you add a member mid-cycle, you’re charged immediately with proration applied. For example, inviting someone on February 15th means you’d pay roughly half the monthly fee (~€17.50) for the remainder of that month. If you then remove that member on February 16th, no refund is issued for the current billing period, but the seat is removed at the start of the next cycle (March 1st), so no further charges apply.

What is the easiest way to try Orq.ai?

You can sign up for a free Developer account to start building with your own data. The free plan includes 50k spans per month, and access to all core platform features with some restrictions — no credit card required. See the comparison table above for full details.

What is a span?

A span is the fundamental unit of observability in Orq.ai. Each span represents a discrete operation in your AI application, such as an model invocation, a tool execution, a retrieval step, or an evaluation. Spans are the building blocks of traces and are automatically created when your application interacts with Orq.ai.

What counts as a span?

A span is recorded whenever a discrete operation occurs within your AI application through Orq.ai. This includes:

  • Each request routed through Orq.ai's AI Gateway (LLM calls, generative model invocations)

  • Individual steps within an agent's execution (tool calls, reasoning steps, retrieval)

  • External spans you send from your own infrastructure via the SDK or API

  • Evaluator and guardrail executions

  • Feedback and scoring events

What is the difference between a span and a trace?

A trace represents a complete end-to-end interaction in your application, such as a user request or an agent invocation. A span is a single step within that trace. For example, when an agent processes a user query, the full interaction is captured as one trace, while each individual operation within it (an LLM call, a tool execution, a retrieval step) is recorded as a separate span. Billing is based on spans.

What is an agent run?

An agent run is recorded each time an agent is invoked through Orq.ai's Agent Runtime. A single agent run may contain multiple spans (LLM calls, tool executions, memory lookups, etc.), giving you full visibility into every step of the agent's behavior.

What does processed data mean?

Processed data is the total volume of data ingested by Orq.ai, including inputs, outputs, prompts, and metadata. It is measured on disk and resets each billing cycle. For example, 100k requests with a 50-page PDF attached will generate far more processed data than 100k single-sentence requests — volume depends on the size of your data, not just the number of requests.

What is a knowledge base, what is a memory store, and what does "2 free" mean?

Knowledge bases and memory stores are both built on top of a vector database but serve different purposes. A knowledge base is fully controlled by the builder — you upload and manage documents, and can expose them on deployments and agents for retrieval-augmented generation (RAG). A memory store is filled dynamically by agents at runtime, such as conversation history or contextual memory. The free plan includes 2 combined — that could be 2 knowledge bases, 2 memory stores, or 1 of each.

What is an agent, and how are agents counted?

An agent in Orq.ai is an AI system you build and run through the Agent Runtime, with support for tools, memory, multi-step reasoning, and orchestration. Agents are counted separately from deployments. The number of agents you can create depends on your plan.

What is a deployment?

A deployment exposes a prompt and model configuration to the Orq.ai API so your application can call it in production. Deployments include versioning, retries, fallbacks, and contextual rules. The free plan includes 3 deployments.

What does a higher rate limit mean in the paid plan?

The paid plan offers higher API rate limits, allowing your applications to handle more concurrent requests. This is important for production workloads with higher traffic. Exact limits depend on your plan.

What are platform APIs and Enterprise APIs?

The platform API is the standard Orq.ai API used to invoke deployments, agents, manage prompts, log traces, and interact with the platform programmatically. The Enterprise API extends this with additional capabilities such as SCIM provisioning, audit logs, and bulk operations for larger organizations with advanced administration needs.

What is a playground?

A playground is a no-code workspace for rapidly iterating on prompts and model configurations. You can compare model and prompt behavior side-by-side, fine-tune parameters, and share configurations with teammates — all before moving to a deployment or experiment.

What does document processing priority mean?

Document processing priority determines how quickly your files are parsed and indexed when uploading documents to a knowledge base. Enterprise customers receive priority processing, meaning their documents are chunked and made searchable faster.

Which plan is right for me?

  • Developer Free: For individual developers or single teams getting started. Includes 50k spans/month and access to core platform features with some restrictions. See the comparison table above for details.

  • Developer Paid: For small teams that need higher limits. Includes 100k spans/month, with additional spans at €7 per 100k. Unlocks higher usage limits and longer retention.

  • Enterprise: For larger organizations with custom requirements such as high data volumes, advanced security needs, on-premise deployment, dedicated support, and role-based access control.

When do I get billed?

You are billed at the end of your monthly billing cycle. All included allowances reset at the start of each new cycle.

What happens if I exceed my included limits?

Any usage beyond your included allowances is billed against metered pricing and appears as separate line items on your invoice. You can review the metered rates for spans, agent runs, and processed data in the pricing table above.

Can I manage or cancel my subscription?

Yes, you can manage or cancel your paid Developer subscription at any time. Changes take effect at the end of your current billing cycle. For Enterprise plans, please contact our team.

Can I use my own private or fine-tuned models?

Yes! All Orq.ai plans support private and fine-tuned models. You can connect your own models through the AI Router alongside 300+ supported models from 20+ providers. For more information, feel free to contact our team.

Does Orq.ai support self-hosted or on-premise deployment?

Yes, self-hosted and on-premise deployment options are available for Enterprise customers. This includes deployment in your own VPC or private cloud to meet stricter compliance and data residency requirements. Contact us to discuss your setup.

What security and compliance standards does Orq.ai meet?

Orq.ai is SOC 2 Type II certified, GDPR compliant, and aligned with the EU AI Act, ensuring enterprise-grade security and data protection. For organizations with advanced security needs, we offer self-hosted, VPC, or hybrid deployment options. Visit our Trust Center at trust.orq.ai for full details.

Can I redline contracts?

Yes, we offer customized contracts for Orq.ai Enterprise customers with a yearly commitment. Please contact us for more details. The Developer plans are designed to be self-serve on our standard terms.

I have security questions, where do I start?

We publish our security documentation and controls publicly. Please visit our Trust Center at trust.orq.ai for full details, or contact our team directly.

How does seat-based billing work when adding or removing team members?

Seats are automatically adjusted: added when you invite a member, and removed (after the current invoice is paid) when you remove one. When you add a member mid-cycle, you’re charged immediately with proration applied. For example, inviting someone on February 15th means you’d pay roughly half the monthly fee (~€17.50) for the remainder of that month. If you then remove that member on February 16th, no refund is issued for the current billing period, but the seat is removed at the start of the next cycle (March 1st), so no further charges apply.

A single control tower for visibility across teams

Book a demo

Explore docs

A single control tower for visibility across teams

Book a demo

Explore docs