Best Dialogflow Alternatives in 2026 for Teams That Want LLM Agents, Not Intents

Quick answer: Google's Dialogflow now sits inside a product called Conversational Agents, and it has no monthly fee. You pay per request: $0.007 for a text request on the classic intent-and-flow edition (Dialogflow CX), $0.012 on the generative Playbooks edition, with voice billed per second. Teams look elsewhere when they want a builder their non-developers can use, a flat monthly price instead of a per-request meter, their own choice of language model, or a chat widget and inbox that arrive ready-made. The best Dialogflow alternatives in 2026 are Botpress (81/100, developer-grade and LLM-native, from $189/mo monthly-billed), Voiceflow (75/100, collaborative design and voice, demo-gated pricing), Chatbase (73/100, answers from your documents, from $40/mo), Typebot (70/100, open-source flows with your own model key, from $39/mo) and Tidio (75/100, website chat plus AI agent, from $29/mo). Dialogflow carries no editorial score here because we have not published its Tier 1 review.

Dialogflow is still a serious product, and the people who should stay on it are easy to describe: engineering teams already living in Google Cloud, with deterministic flows they have tuned for years, voice or telephony traffic, and a per-request bill they are comfortable forecasting. Google's own editions guide positions Dialogflow CX as "suitable for large or complex applications" and the older Dialogflow ES as suitable for "small to medium, moderately complex applications."

We have not published a Tier 1 review of Dialogflow, so it carries no editorial score and this page makes no claim about its conversation quality against the five platforms we have tested; a full Dialogflow review is on our hands-on queue. What this page rests on instead is Google Cloud's own pricing page and editions documentation, both read on 9 October 2026. Alternative scores and prices come from each platform's published review ledger at monthly-billed rates, per our pricing methodology. Who did this work is on our editorial team page.

One scope note. This page is for the team choosing a platform for a business chatbot: a website widget, WhatsApp, a support deflection layer. If you are building a phone-call IVR replacement on Google Cloud, Dialogflow's voice pricing and telephony ties are a different decision; our IVR to AI voice migration guide and best voice AI platform ranking cover that. If you are choosing from scratch, start with our best chatbot builder ranking.

Why you are looking for a Dialogflow alternative

Four reasons come up most. Each traces to a line in Google's own documentation, not to a complaint about the NLU.

The product has more than one name and more than one edition

Google's pricing page now describes Conversational Agents as an umbrella with two editions: Flows, "also known as Dialogflow CX," which are deterministic agents built from intents and flows, and Playbooks, generative agents built from natural-language instructions with data stores, generative fallbacks and generators. Dialogflow ES, the older edition, has its own separate pricing page, and the editions doc lists a "Conversational Agents console" as the new interface replacing the Dialogflow CX console. A buyer comparing prices therefore has to decide between three things before pricing anything: ES, CX Flows, or Playbooks. That is not a flaw in the engineering. It is a cost in evaluation time that a single-product tool does not impose.

The bill is per request, and the number of requests is up to your design

Chat is billed per request: $0.007 for Flows, $0.012 for Playbooks, and ES Essentials is $0.002 per request. Google defines a request as any API call to the platform, whether made directly, through an integration or from the console, and adds that "the number of requests needed to complete a task varies with agent design." So the price per conversation is not a number you can read off a page; it depends on how many calls your design makes per turn.

Here is an illustration with invented numbers, not a benchmark. Take 1,000 conversations a month at 8 turns each, and assume one request per turn: 8,000 requests. On Flows that is $56 a month; on Playbooks it is $96. If your design makes two calls per turn, double both. At that scale Dialogflow can be cheaper than any flat plan below, which is a reason to stay. At 20 times that volume the meter is $1,120 to $1,920 a month before voice, hosting or engineering time, and a flat plan can win.

Generative features are billed at a different rate, and mixing them reprices the turn

When a Flow calls a Playbook, Data Store, Generator or generative fallback, that turn is billed at the Playbook rate; if a Playbook calls a Flow in any turn, Google's page says all turns are billed as Playbook requests. Data store storage is free up to 10 GiB a month, then $5.00 per GiB. Voice is billed per second, rounded up, and the full generated audio is billed even if the caller interrupts it. None of this is hidden, but forecasting it takes a spreadsheet. Our pricing for AI chatbots guide explains why per-event meters behave this way, and our token and cost estimator guide shows how to model one.

The builder is built for developers

The editions doc describes ES as mostly text forms and CX as a state-machine model with pages as nodes and visual conversation graphs; it lists building a complex agent as moderate difficulty in CX. That is a capable tool for an engineering team, but it is not a marketing-team tool. If the people who will maintain the bot are support leads rather than developers, or if you need the bot live on a website and WhatsApp this week with a shared inbox behind it, the platforms below ship more of that surrounding product.

And the reasons to stay. There is no monthly minimum, so a low-volume bot can cost a few dollars. Free trial credits are $600 for Flows and $1,000 for Playbooks, but they expire after 12 months, so test early. CX allows up to 1,000 agents per project against one for ES, which matters to agencies and large organizations. Design-time requests, such as using the visual builder, are free. And if your data and identity already live in Google Cloud, consolidation is a real benefit that none of the alternatives can offer.

How the top Dialogflow alternatives compare

Dialogflow figures come from Google Cloud's pricing page and editions documentation, read 9 October 2026. All alternative prices are monthly-billed rates carried from each platform's review ledger.

The 5 best Dialogflow alternatives

The list runs in fit order, not score order. Readers who arrive from Dialogflow usually have some engineering background, so the closest-in-spirit tools come first, and Tidio, which scores higher than two of the others, sits last because it asks the least of you and replaces the least of Dialogflow.

1. Botpress — The nearest thing to Dialogflow for teams that want LLM agents

Botpress is the strongest all-round answer for a team that picked Dialogflow because it wanted control. It combines a node-based Studio, a TypeScript development kit that deploys into the same runtime, a 200-plus integration hub, and an LLM-orchestration layer, where the model decides which tools, knowledge bases or sub-flows to call. Our reviewer measured 14 minutes from signup to a working FAQ bot on the native web chat channel, and an 86% first-answer accuracy on 15 questions against a five-PDF knowledge base, with a 9% hallucination rate on the Plus tier. It supports OpenAI, Anthropic, Groq and Hugging Face models, and speaks the Model Context Protocol in both directions (see our MCP glossary entry).

The difference from Dialogflow is the meter and the sticker price. Botpress bills conversations with a bundled AI allowance and recharges AI usage beyond it at cost, which is easier to forecast than a count of API calls but starts at $189 a month rather than at zero. Our review scores pricing 60/100 and says the learning curve is real: G2 reviewers split between ease of use and a steep production learning curve.

Choose Botpress over the others when: you have developers and want the platform to be the runtime, not just the builder. Skip it if: the bot will handle a few hundred conversations a month and nobody will use the code surface, where Dialogflow's pay-as-you-go or a $40 plan costs far less.

Editorial score: 81/100 · Read our full Botpress review → · See also Botpress alternatives and Botpress vs Voiceflow

2. Voiceflow — When designers and support leads build alongside engineers

Voiceflow is the alternative closest to Dialogflow CX's flow-and-state way of thinking, with a more designer-friendly canvas. Its agents are a hybrid: explicit nodes (capture, intent, speak, API call, function) with the model invoked at specific points, which keeps conversations predictable and easy for non-developers to audit. In our test the agent used the knowledge base in 89% of the turns where it applied, and all three function nodes fired on the right intent. Voice is a first-class channel: it can provision a US or Canadian phone number directly and also accepts Twilio, Vonage and Telnyx numbers. It supports multiple model providers, including bring-your-own, and since September 2026 offers live agent handoff on all plans.

The catch is the one that sends many small teams elsewhere: you cannot learn the price without a demo or an account. That is the opposite of Dialogflow's published per-request rates. Our review also notes the platform's US enterprise-CX skew.

Choose Voiceflow over the others when: designers or support leads own the conversation and voice matters. Skip it if: you need a price before you invest an afternoon, or want a messenger-marketing tool.

Editorial score: 75/100 · Read our full Voiceflow review → · See also Voiceflow alternatives and Botpress vs Voiceflow

3. Chatbase — When the real goal is "answer questions from our docs"

If your Dialogflow project is mostly an FAQ bot, the intent-and-training-phrase model is more work than the job needs. Chatbase drops it: you upload files, crawl URLs or add Q&A pairs, and a retrieval-based agent (RAG) answers from them. Our reviewer ingested a five-PDF, 80-page set in four minutes, the fastest we have measured, and built an AI Action that looked up a Stripe subscription in 17 minutes. WhatsApp, Messenger, Instagram and Slack are included from Hobby, and handover to a human, rated 4/5 in our review, carries the conversation history across.

The limits are structural. Chatbase is a support-answering product, not a flow designer, so deterministic multi-step journeys are not its strength, and the model is vendor-managed: our review found no documented path to your own API key. Credits also scale with the model you choose, from 1 credit per reply on the recommended low-cost options up to 5 on the most expensive, so the same plan lasts very different lengths of time. Extra agents cost $25 a month each.

Choose Chatbase over the others when: your bot's job is answering from documents and you want it live today. Skip it if: you need branching flows, your own LLM key, or an open-source license.

Editorial score: 73/100 · Read our full Chatbase review → · See also Chatbase alternatives and Botpress vs Chatbase

4. Typebot — Open-source flows with your own model key

Typebot suits the Dialogflow user who valued ownership. It is an open-source builder (about 10,500 GitHub stars as of 7 October 2026) that you can self-host under the Functional Source License, which permits self-hosting and modification but prohibits using the code to compete with its cloud service. Its AI blocks connect to OpenAI, Anthropic, Mistral and others using your own API key, so model spend is billed by the provider directly. It is the only option here that is both open source and bills model spend directly to your own provider key, and our review calls bring-your-own-key the structurally rarest capability in our batch at small-business pricing. Our reviewer had a website FAQ bot live in 11 minutes. See our BYOLLM glossary entry and the BYOLLM chatbot guide for why this matters.

The trade is channel depth and polish. Typebot is web-first; WhatsApp needs the $89 Pro plan and Typebot is not a Meta Business Solution Provider, so template approval took 58 hours in our test, against 28 on a BSP-based WhatsApp specialist. Human handover runs through integrations such as the Chatwoot and Zendesk blocks our review documents, not a shared inbox like Tidio's, and billing is monthly-only.

Choose Typebot over the others when: you want ownership, own-key billing and web-first flows. Skip it if: WhatsApp is the main channel or non-developers will maintain it.

Editorial score: 70/100 · Read our full Typebot review → · See also Typebot alternatives and Botpress vs Typebot

5. Tidio — The ready-made option: website chat, inbox and an AI agent

Tidio is the opposite end of the spectrum from Dialogflow, and a reasonable answer for a small business that does not need Google Cloud. A chat widget, shared inbox and ticketing came first and the bots second. Lyro, its AI agent, runs on Anthropic's Claude and answers from your knowledge sources. Our reviewer rated its inbox and handover 4.5/5, and a widget with a working quick-reply flow took nine minutes. Free, Starter and Growth each include ten seats, and a native Shopify connector shows order status in the chat.

What you give up is the engineering surface. There is no per-request API model to design, no flow state machine of Dialogflow CX's depth, and no voice stack. Tidio meters conversations (50 on Free, 100 on Starter, up to 2,000 on Growth at $59), and Lyro's 50 AI conversations on Free and Starter are a one-off allowance, so ongoing AI answering means Growth.

Choose Tidio over the others when: you need live website chat with an AI agent now and nobody on staff writes code. Skip it if: you are building complex branching journeys or a voice agent.

Editorial score: 75/100 · Read our full Tidio review → · See also Tidio alternatives and best website chatbot

Also compared, not scored. Rasa, Amazon Lex and Microsoft Copilot Studio come up in almost every Dialogflow switching search, and Google's own Playbooks edition is the in-house route to a generative agent. We have not reviewed them, so they are not ranked here. The first two are developer frameworks and services in the same spirit as Dialogflow; if you want that class of tool, our best developer chatbot and best open source chatbot pages are the closest we have, and our best AI agent ranking covers agent platforms more broadly.

How to choose your Dialogflow replacement

Start from who will maintain the bot and what it will cost per month at your volume.

Decide whether you have an intent problem or a documents problem

If the bot answers questions that already exist in help articles, an intent model is the long way round. A retrieval-based agent (Chatbase, Tidio's Lyro, or Botpress with a knowledge base) gets you there faster; see our chatbot knowledge base guide and build a RAG chatbot guide. If the bot must complete multi-step tasks with strict rules, such as collecting slots and validating them, keep a flow tool: Botpress, Voiceflow or Typebot. Our no-code vs low-code chatbot guide helps place your team.

Put Dialogflow's meter next to a flat plan

Count last month's turns and the API calls per turn, multiply by $0.007 or $0.012, and compare with the monthly plan of the tool you are considering. The break-even is where a plan like Tidio Starter at $29 or Chatbase Hobby at $40 beats your Google Cloud bill, and it moves with your design. Our LLM API cost calculator and bot vs human cost calculator give you the rest of the sum.

Check handover before you migrate

Write down how a conversation reaches a human today, then test the same case on a trial of each candidate. Our human handoff glossary entry and handoff design guide cover what a good transfer carries, and the vendor evaluation checklist turns it into a scorecard.

Run both, then cut over

Export your intents, training phrases and entities, and rebuild the top twenty by volume first. Run the new bot on one page or channel while Dialogflow keeps the rest, and compare fallback rate on both for two weeks. Our 14-day migration playbook has the sequence. One Dialogflow-specific check: use your trial credits ($600 for Flows or $1,000 for Playbooks, expiring after 12 months) to test before you commit elsewhere.

Frequently asked questions

Author: Chatbotscape Editorial Methodology version: 2026-Q2 (How we test) Last verified: 9 October 2026 (Google Cloud Conversational Agents pricing and Dialogflow editions documentation read the same day) Next review: 10 January 2027 Affiliate disclosure: Chatbotscape earns commission on paid sign-ups via affiliate links on this page. This does not influence our editorial scores — see our full affiliate disclosure policy.