Dialogflow Pricing: Understanding Costs for ES and CX (2026)

Pricing overview

Dialogflow's pricing model is usage-based, differentiating costs between its two primary editions: Dialogflow ES (Essentials) and Dialogflow CX (Customer Experience). Users are charged based on the number of requests made to the agent, the duration of audio processed, and the amount of data stored, among other factors. Both editions include a free tier designed to cover initial development and testing, with charges applying once these free limits are exceeded. The specific rates vary significantly between ES and CX due to their differing capabilities and target use cases, with CX generally designed for more complex, enterprise-grade conversational AI applications and therefore having a higher per-unit cost Google Cloud Dialogflow pricing details.

Plans and tiers

Dialogflow does not offer traditional subscription plans with fixed monthly fees. Instead, its pricing operates on a pay-as-you-go model, where costs scale directly with usage. The primary distinction in pricing is between Dialogflow ES and Dialogflow CX.

Dialogflow ES Pricing

Dialogflow ES is generally suited for simpler agents and small to medium-scale applications. Its pricing is structured around:

• Text Queries: Billed per 1,000 text requests.
• Audio Input/Output: Billed per minute of audio processed for both speech-to-text and text-to-speech.
• Data Storage: Charges apply for the amount of agent data stored.
• Knowledge Connector: Additional charges for knowledge base queries.

Dialogflow CX Pricing

Dialogflow CX is designed for large-scale, enterprise-grade virtual agents requiring complex conversational flows and visual design tools. Its pricing reflects its advanced capabilities:

• Session Charges: Billed per session, which represents a conversational interaction between an end-user and a Dialogflow agent. A session is defined as a continuous exchange of turns with a maximum duration of 30 minutes, regardless of the number of turns within that period Dialogflow CX session pricing.
• Audio Input/Output: Similar to ES, billed per minute of audio processed.
• Data Storage: Costs for storing agent data.
• Interaction Logging: Charges for storing conversational logs.
• Flows, Pages, and Intents: While not directly billed per unit, the complexity enabled by these features contributes to the value proposition of CX, justifying the higher session costs compared to ES.

Here is a comparison of the general pricing structure for both editions:

Free tier and limits

Both Dialogflow ES and Dialogflow CX offer a free tier, allowing developers to experiment and build conversational agents without incurring immediate costs. These free tiers are subject to specific monthly usage limits.

Dialogflow ES Free Tier

The Dialogflow ES free tier includes:

• Text Queries: Up to 30,000 text requests per month.
• Audio Input/Output: Up to 100 minutes of audio processing (both speech-to-text and text-to-speech) per month.
• Data Storage: A limited amount of free data storage.

These allowances are typically sufficient for developing and testing small to medium-sized agents or for initial proof-of-concept projects. Usage beyond these limits is billed at the standard pay-as-you-go rates Dialogflow free tier specifics.

Dialogflow CX Free Tier

The Dialogflow CX free tier provides a more generous allowance to support the development of more complex agents:

• Sessions: Up to 100 sessions per month.
• Audio Input/Output: Up to 100 minutes of audio processing per month.

The CX free tier is designed to enable extensive testing of complex conversational flows. Once the 100-session limit or 100-minute audio limit is reached within a month, standard CX session and audio processing rates apply.

It's important to monitor usage within the Google Cloud console to avoid unexpected charges, especially when transitioning from the free tier to paid usage. The free tier is a critical component for developers to evaluate the platform before committing to larger-scale deployments.

Real-world cost examples

Estimating real-world costs for Dialogflow depends heavily on the chosen edition, the volume of interactions, and the complexity of the agent. Here are a few illustrative scenarios:

Example 1: Small Customer Service Chatbot (Dialogflow ES)

• Scenario: A small business deploys a Dialogflow ES chatbot on their website to answer frequently asked questions. The chatbot handles approximately 25,000 text queries per month and uses minimal audio processing (e.g., for internal testing, not public IVR).
• Cost Breakdown:
  • Text Queries: 25,000 queries. Since the free tier covers 30,000 queries, this usage would largely fall within the free allowance.
  • Audio Processing: Negligible.
  • Estimated Monthly Cost: Close to $0, assuming usage stays within the free tier for text queries and audio Google Cloud pricing rates.
• Consideration: If the chatbot scaled to 50,000 text queries, the additional 20,000 queries would be billed at the standard rate (e.g., $0.002 per query, or $2.00 per 1,000 queries, resulting in an extra $40.00).

Example 2: Medium-Sized IVR System (Dialogflow ES with Audio)

• Scenario: A company uses Dialogflow ES to power a moderately busy interactive voice response (IVR) system. It handles 100,000 text queries and 5,000 minutes of combined speech-to-text and text-to-speech audio processing per month.
• Cost Breakdown:
  • Text Queries: 100,000 queries. After the 30,000 free queries, 70,000 queries are billed. At a hypothetical rate of $0.002 per query (or $2.00 per 1,000 queries), this would be $140.00.
  • Audio Processing: 5,000 minutes. After the 100 free minutes, 4,900 minutes are billed. At a hypothetical rate of $0.06 per minute, this would be $294.00.
  • Estimated Monthly Cost: Approximately $434.00.

Example 3: Enterprise Omnichannel Virtual Agent (Dialogflow CX)

• Scenario: A large enterprise deploys a Dialogflow CX agent across multiple channels (web, mobile, voice) for complex customer support. It handles 50,000 customer sessions and 10,000 minutes of audio processing per month.
• Cost Breakdown:
  • Sessions: 50,000 sessions. After the 100 free sessions, 49,900 sessions are billed. At a hypothetical rate of $0.02 per session, this would be $998.00.
  • Audio Processing: 10,000 minutes. After the 100 free minutes, 9,900 minutes are billed. At a hypothetical rate of $0.06 per minute, this would be $594.00.
  • Estimated Monthly Cost: Approximately $1,592.00.

These examples use hypothetical rates for illustration; actual rates are available on the official Google Cloud Dialogflow pricing page official Dialogflow pricing. It's crucial to factor in that Dialogflow pricing can vary by region and may include additional costs for premium features, knowledge connectors beyond basic usage, or specific integrations.

How the pricing compares

Dialogflow's pricing model, particularly its pay-as-you-go structure with distinct ES and CX editions, is comparable to other major cloud-based conversational AI platforms. Competitors like Amazon Lex and Microsoft Bot Framework (which can integrate with Azure Cognitive Services) also employ usage-based billing models, typically charging per request, message, or session, and for associated services like speech-to-text and text-to-speech.

• Amazon Lex: Charges per text request and per minute of speech input. It also offers a free tier for initial usage. The cost structure is broadly similar to Dialogflow ES in its per-request model Amazon Lex pricing details.
• Microsoft Bot Framework (Azure Bot Service + Cognitive Services): While the Bot Framework itself is often free or low-cost, the actual conversational AI capabilities (like LUIS for NLU or Azure Speech Service) are billed separately as part of Azure Cognitive Services. These typically follow a per-transaction or per-character/minute model, analogous to Dialogflow's query and audio processing charges Azure Cognitive Services Language pricing.
• IBM Watson Assistant: Offers a tiered pricing model that includes a free tier and then moves to paid plans based on monthly active users (MAUs) or API calls, with separate charges for premium features. This MAU model can differ from Dialogflow's per-session or per-query approach, potentially offering more predictable costs for known user bases but less granularity for highly variable usage IBM Watson Assistant pricing plans.

The choice between these platforms often comes down to specific feature requirements, ecosystem preference (e.g., existing Google Cloud, AWS, or Azure investments), and the complexity of the desired conversational experience. Dialogflow CX, with its session-based billing and advanced flow management, tends to be positioned for more sophisticated enterprise needs, potentially at a higher per-unit cost than simpler alternatives or Dialogflow ES, reflecting its enhanced capabilities for managing intricate multi-turn conversations and state management. Developers should carefully evaluate the detailed pricing pages of each vendor against their projected usage patterns to determine the most cost-effective solution for their specific use case.