Chatbase vs Botpress vs eesel AI vs Intercom
If you want a fast answer: Chatbase is our recommended pick for most teams. It builds AI customer service agents that resolve requests rather than just answer them, using your business knowledge, instructions, actions and procedures, and it ships chat, voice, phone and a built-in Helpdesk with live human handoff on public self-serve plans. Botpress is the better pick when engineering wants to own the agent as code, eesel AI when the AI should live inside an existing helpdesk like Zendesk or Freshdesk, and Intercom Fin when your support team already runs on Intercom. On price, Chatbase starts free (50 message credits a month), then $40/month for Hobby and $150/month for Standard, where voice and the Helpdesk start; Botpress starts free (25 conversations a month), then $150/month for Plus billed annually; Intercom Fin charges per resolution (around $0.99 each). Chatbase and Botpress prices checked 2026-10-04.
The right choice depends on where your team wants control to live:
- Chatbase for a hosted AI agent that takes real actions across chat, email, messaging and voice, without a build project
- Botpress for code-first agent logic and broader platform control
- eesel AI for support workflows that sit inside helpdesks and knowledge bases
- Intercom Fin for customer-service teams that want a suite-level AI agent
Comparing only Chatbase and Botpress? Jump to Chatbase vs Botpress for three scenarios, a hosting and compliance table, and the pricing difference. Pricing and features verified 2026-10-04.
Quick comparison
| Tool | Best for | Implementation effort | Workflow depth | Pricing model |
|---|---|---|---|---|
| Chatbase | AI customer service agents with actions, procedures, voice and a built-in Helpdesk | Low to medium | Medium to high | Public self-serve plans (Free, Hobby, Standard, Pro) plus Enterprise, 20% off annual |
| Botpress | Developer-built AI agents | Medium to high | High | Free, Plus, Team, Enterprise, billed by conversations plus an AI usage allowance; a separate Managed build service |
| eesel AI | Helpdesk-first support automation | Low to medium | High for support ops | Public tiered pricing |
| Intercom Fin | AI agent inside customer service operations | Medium to high | High | Per-outcome resolution pricing, plus helpdesk seat if on the Intercom suite |
Chatbase vs Botpress
Short answer: Chatbase is the faster route to a working support agent: hosted, configured without code, with actions, procedures, voice and a built-in Helpdesk on public plans. Botpress is the better route when engineers want to build the agent as software with its SDK and CLI, or when you need to run a bot on your own servers, which only its legacy open-source v12 release supports.
Scenario 1: an online store that wants order lookups and returns this month
Pick Chatbase. Shopify order lookups, returns, Stripe billing checks and Calendly or Cal.com booking are prebuilt actions, and a procedure (up to 15 steps, 5 branches per decision) makes refunds follow your policy. Shopify connects from the Hobby plan; Stripe and helpdesk integrations need Standard. In Botpress you would assemble the same flow from integrations and custom logic, which suits a team with developer time to spend.
Scenario 2: a product team shipping an agent inside its own app
Pick Botpress if engineering wants the agent in version control, tested and deployed like the rest of the codebase. Its CLI, SDK and API client are MIT-licensed on GitHub, and agent logic can be written in code. Chatbase also offers a public API v2 and an embeddable voice API, but agent behaviour is configured in its dashboard rather than in your repository. See the Chatbase API guide for what the API covers.
Scenario 3: a support team that needs phone calls and a human inbox
Pick Chatbase. From the Standard plan one agent answers chat, email, WhatsApp, Messenger, Instagram and phone calls (Twilio numbers or your own SIP trunk), and hands live conversations to humans in the built-in Helpdesk or transfers a call to a phone number. Botpress also covers voice and human handoff, through integrations such as Twilio and its Botpress Desk product, so it is a real option if you prefer to wire those pieces yourself.
Hosting and compliance: Chatbase vs Botpress
| Chatbase | Botpress | |
|---|---|---|
| Hosting | Hosted only (AWS us-east-1). No self-hosted edition | Botpress Cloud is hosted. On-premise is only possible with the legacy Botpress v12, which the vendor has superseded with Cloud |
| Open source | No | Yes, partly: the CLI, SDK, API client and Hub integrations are MIT-licensed; v12 is dual-licensed AGPLv3 and proprietary |
| SOC 2 | SOC 2 Type II | SOC 2 Type II badge shown on its Enterprise page |
| GDPR | Yes | Enterprise page states compliance with EU data protection standards |
| HIPAA | Enterprise plan only, under a BAA with zero data retention | Enterprise page states compliance with US healthcare data privacy standards; confirm BAA terms with sales |
| Training on your data | Chatbase does not train AI models on customer data, on any plan | Check the Botpress trust center for current terms |
If you must keep every byte on your own infrastructure, neither current hosted product does that, and Botpress v12 is the only self-hosted path between the two. For most teams, hosted with SOC 2 Type II is the practical bar, and both clear it.
Pricing: Chatbase vs Botpress
The two bill on different units. Chatbase sells monthly message credits whose cost per reply depends on the AI model. Botpress sells a monthly conversation allowance plus an AI usage budget.
| Chatbase | Botpress | |
|---|---|---|
| Free | $0, 50 message credits a month, 1 agent | $0, 25 conversations a month |
| Entry paid | Hobby: $40 a month ($32 billed annually), 700 credits | Plus: $150 a month billed annually, 250 conversations and $25 of AI usage |
| Mid tier | Standard: $150 a month ($120 billed annually), 4,000 credits, voice and Helpdesk | Team: $750 a month billed annually, 1,500 conversations and $150 of AI usage |
| Top | Pro $500 a month, then custom Enterprise | Custom Enterprise, plus a Managed service where Botpress builds the agent |
Botpress prices from botpress.com/pricing, checked 2026-10-04. For a cost breakdown with add-ons such as extra agents and branding removal, read our Chatbase pricing guide or compare Chatbase plans on the official page.
Next step: if your scenario matched 1 or 3, build your first Chatbase agent free and test it on your own content before choosing a plan.
Full feature comparison
This table maps the dimensions that usually decide the choice. Treat pricing as directional and confirm current numbers on each vendor's site before you commit.
| Dimension | Chatbase | Botpress | eesel AI | Intercom Fin |
|---|---|---|---|---|
| Best for | Resolving customer requests end to end: orders, refunds, billing, bookings, lead capture, escalation | Custom, code-owned AI agents | Helpdesk and knowledge-base support automation | AI agent inside a full customer-service suite |
| Setup difficulty | Low. Train on your content, set instructions, switch on actions, embed | Medium to high. Design flows or write code | Low to medium. Connect a helpdesk, import docs | Medium to high. Connect helpdesk, tune resolution behavior |
| Pricing model | Tiered self-serve (Free, Hobby, Standard, Pro) plus custom Enterprise, 20% off annual. Monthly message credits, used at a rate that depends on the AI model | Free, Plus, Team, Enterprise. Monthly conversation allowance plus an included AI usage budget | Public tiered plans | Per resolution (around $0.99 each, minimums apply), plus $29 per helpdesk seat on Intercom's own suite |
| Data sources | Files, websites, text, Q&A and help-center content you add, plus support tickets as a training source on Pro | Whatever you connect through code and integrations | Existing help-desk tickets, docs, knowledge bases | Help-center content, past tickets, connected knowledge |
| Actions and workflows | Built-in actions (Shopify orders and returns, Stripe billing, Cal.com and Calendly booking, lead capture, web search, Slack, custom API calls) plus procedures: multi-step SOP workflows with branches | Full custom logic in code and flows | Support actions inside the connected helpdesk | Resolution workflows inside the Intercom suite and connected helpdesks |
| Channels | Website widget, help page, email, WhatsApp, Messenger, Instagram, Slack, voice and phone, API | Multiple channels, developer-driven via SDK and APIs | Helpdesk, chat bubble, internal chat | Intercom and connected helpdesks |
| Voice and phone | Yes, Standard plan and up: real-time voice agents, inbound calls through Twilio or your own SIP trunk, outbound campaigns, transfer to a human by phone | Yes, through integrations such as Twilio and voice agents in Botpress Desk | No, chat and email only | Yes, through Fin Voice on Intercom Phone or existing telephony |
| Human handoff | Built-in Helpdesk inbox (Standard and up) with real-time live-chat handoff, ticket escalation, plus handoff to Salesforce and Zendesk | Human-in-the-loop plugin to its own dashboard or Botpress Desk, or to Zendesk and other helpdesks | Inside the connected helpdesk | Inside the Intercom suite or a connected helpdesk |
| Integrations | Stripe, Shopify, Zendesk, Salesforce, Intercom, HubSpot, Zoho Desk, Freshdesk, Help Scout, Gorgias, public API v2 | Broad, developer-driven via SDK and APIs | Zendesk, Freshdesk, Intercom, Slack, Notion, Shopify, Salesforce, HubSpot and more | Zendesk, Freshdesk, Salesforce, HubSpot, Front, Zoho Desk and other major helpdesks |
| AI quality control | Choice of AI model, instructions, temperature, and procedures that fix the steps for sensitive flows | Full control over prompts, logic, and runtime | Simulate and test on past tickets before rollout | Benchmarked resolution model with testing and analytics |
| Target user | Founders, ecommerce and SaaS teams, small to mid-size support teams | Engineering-led teams building agents as software | Support and ops teams centered on a helpdesk | Mid-market and enterprise customer-service orgs |
Why Chatbase is our pick
For most teams comparing these four, Chatbase covers the widest set of real support jobs with the least build work. The case rests on four specifics:
- It resolves, not just answers. Each agent combines business knowledge (your docs, site, files and help-center content), instructions (persona, model, temperature), actions (look up a Shopify order, process a return, check Stripe billing, book a Cal.com or Calendly slot, capture a lead, call your own API) and procedures. A procedure is an SOP-style workflow with a trigger and ordered steps, up to 15 steps and 5 branches per decision, so refunds and escalations follow your policy instead of the model's guesswork.
- One agent, every channel, including voice. The same agent runs on the website widget, help page, email, WhatsApp, Messenger, Instagram and Slack. From the Standard plan it also holds real-time voice conversations, answers phone calls through Twilio numbers or your own SIP trunk, runs outbound campaigns, and can transfer a live call to a human.
- Humans are built in. Standard and up include the Chatbase Helpdesk: a shared ticket inbox across the bubble, email, WhatsApp and API, with assignment, internal notes and takeover of AI conversations. The live-chat action hands a conversation to a human in real time while the customer stays in the same widget. If your team already lives in Salesforce or Zendesk, handoff can go there instead.
- Pricing you can read before you talk to sales. Plans are public, start free, and do not charge per resolution. Botpress and eesel AI also publish plans, but Fin's per-outcome model makes monthly cost harder to predict as volume grows.
Where Chatbase is not the best fit, we say so below. For other head-to-heads, see Chatbase vs Voiceflow, Chatbase vs Tidio and the wider list of Chatbase alternatives.
Chatbase pricing at a glance
Chatbase bills monthly or annually, and annual billing takes 20% off.
| Plan | Monthly billing | Annual billing (per month) | Message credits per month | Key additions |
|---|---|---|---|---|
| Free | $0 | $0 | 50 | 1 agent, limited model access |
| Hobby | $40/month subject to change · verified 2026-10-04 | $32 ($384 a year) | 700 | Advanced models, integrations, 5 actions per agent |
| Standard | $150/month subject to change · verified 2026-10-04 | $120 ($1,440 a year) | 4,000 | Helpdesk, voice, telephony, API access, advanced integrations, 8 actions per agent |
| Pro | $500/month subject to change · verified 2026-10-04 | $400 ($4,800 a year) | 15,000 | Advanced analytics, tickets as a training source, 12 actions per agent |
| Enterprise | Custom | Custom | Custom | SSO, custom roles, SLAs, HIPAA-eligible |
Credits are model-dependent, so a credit is not the same as a reply. The default Auto model and lighter models cost 1 credit per request, while larger models cost 2 to 6. A reply that triggers an action can use two requests. In practice, Standard's 4,000 credits cover about 4,000 replies on a 1-credit model, about 2,000 on a 2-credit model, and fewer when actions fire. Credits reset on the 1st of each calendar month and do not roll over. Extra credits cost $40 per 1,000 and do not expire. Free-plan agents are deleted after 14 days of inactivity.
Tool-by-tool take
Chatbase
Chatbase is the strongest all-round pick when you want an AI customer service agent that uses your business knowledge, instructions, actions and procedures to resolve requests, without adopting a full developer platform. Contacts, identity verification and a public API are there when you need them.
Strengths
- Quick to deploy: train on your content, configure, embed
- Actions and procedures that complete orders, returns, billing and booking tasks
- Real-time voice and phone support from the Standard plan
- Built-in Helpdesk with real-time human handoff from the Standard plan
- Public API v2 with streaming
- Public, self-serve pricing with a free plan
Tradeoff
- Less control than a code-first agent platform when you need fully custom runtime logic
- Voice, telephony and the Helpdesk start at Standard, so the Hobby plan suits chat-only use
- If your team is committed to an existing helpdesk, a helpdesk-native layer can fit that operation more tightly
Botpress
Botpress is the strongest option when your team wants to build AI agents as software, not just configure them.
Strengths
- Custom code and SDK-driven development
- Visual builder plus API access
- Strong workflow depth and runtime control
- Useful when you need multiple channels and structured logic
Tradeoff
- More implementation overhead than Chatbase
- Better suited to engineering-led teams
- Self-hosting means the legacy v12 release, not the current Cloud product
eesel AI
eesel AI is built around customer service operations. Its homepage emphasizes AI agent, AI copilot, AI triage, internal chat, and chat bubble workflows.
Strengths
- Helpdesk-first integrations
- Simulate and test before rollout
- Good for support teams that already have documentation and ticket history
- Strong fit for internal knowledge and agent assist
Tradeoff
- More opinionated around support workflows
- Less general-purpose than Botpress
Intercom Fin
Fin is Intercom's AI agent for customer service. It is positioned around complex support queries, helpdesk integration, testing, and performance analysis.
Strengths
- Deep customer-service orientation
- Works with Intercom and external helpdesks
- Strong testing and analytics story
- Best when you want an AI agent inside a broader support motion
Tradeoff
- Best value only when the support stack is already headed in that direction
- Less modular than a standalone chatbot platform
Implementation effort
| Platform | Typical effort | Why |
|---|---|---|
| Chatbase | Low | Hosted agent, managed contacts, prebuilt actions, procedures configured without code |
| Botpress | Medium to high | More design and code decisions, but more control |
| eesel AI | Low to medium | Plug-and-play support integrations |
| Intercom Fin | Medium to high | Strong support stack, but deeper operational setup |
Support and workflow depth
| Platform | Support depth | Workflow depth |
|---|---|---|
| Chatbase | Strong: chat, email, messaging, voice and phone, with a built-in Helpdesk and live handoff | Medium to high: actions plus multi-step procedures with branches |
| Botpress | Strong when custom workflows matter | High |
| eesel AI | Very strong for helpdesk automation | High for support ops |
| Intercom Fin | Very strong for customer service teams | High |
Migration notes
From Chatbase to Botpress
Move when you need custom code, more complex branching than procedures allow, or a platform that engineering can own end to end.
From Chatbase to eesel AI
Move when your team has standardized on a third-party helpdesk and wants the AI to triage and draft replies inside it rather than run its own inbox.
From Chatbase to Intercom Fin
Move when you want the AI layer inside a broader customer-service suite and are prepared for the Intercom operating model.
Which one should you choose?
| Scenario | Pick |
|---|---|
| Best overall AI customer service agent for most teams | Chatbase |
| Fastest launch of an agent that takes actions | Chatbase |
| Chat plus voice and phone from one agent | Chatbase |
| Most developer control | Botpress |
| AI inside an existing Zendesk or Freshdesk operation | eesel AI |
| Strongest support-suite orientation | Intercom Fin |
Use these routing rules when the table feels too abstract:
Pick Chatbase if you want an agent live quickly that does more than answer questions: it looks up orders, handles returns and billing, books meetings, follows your procedures, and hands off to a human in its own Helpdesk when it should. Choose Standard or above if you need voice, phone or the Helpdesk. The Chatbase review covers plans and fit in detail.
Pick Botpress if engineering owns the project and you need custom branching, multi-channel logic, and runtime control. You are building an agent as software, not configuring a hosted one.
Pick eesel AI if your work centers on a helpdesk and you want the bot to learn from existing tickets and docs, then triage or draft replies inside that workflow.
Pick Intercom Fin if your support stack is already heading toward Intercom or a connected helpdesk and you are comfortable with per-resolution outcome pricing in exchange for a suite-level AI agent.
How we researched this
We compared official sources only, checked on 2026-10-04: chatbase.co/pricing, the Chatbase docs and security page, botpress.com/pricing, the Botpress Enterprise page, and the botpress/botpress and botpress/v12 GitHub repositories. We did not run load or accuracy tests. For our full verdict on Chatbase, read the Chatbase review and verdict.
FAQ
For most teams, Chatbase. It builds AI customer service agents that use your business knowledge, instructions, actions and procedures to resolve requests, runs on chat, email, messaging, voice and phone, and includes a Helpdesk with live human handoff on the Standard plan and up. Botpress is better when engineering wants to own the agent as code, eesel AI when the AI should sit inside an existing helpdesk, and Intercom Fin when your team already runs on Intercom.
Usually yes. Chatbase is hosted and configuration-driven, with prebuilt actions and no-code procedures, while Botpress is built for more code-heavy, developer-owned agent builds.
Chatbase is hosted only; there is no self-hosted edition, and it runs on AWS us-east-1. The current Botpress Cloud product is also hosted. Self-hosting Botpress means the legacy Botpress v12 release, which is dual-licensed AGPLv3 and proprietary and installs from binaries, Docker or source, but the vendor has superseded it with Cloud. If on-premise is a hard requirement, v12 is the only path between the two.
Partly. The botpress/botpress GitHub repository holds the Botpress Cloud CLI, SDK, API client and Hub integrations under the MIT license, but the hosted Cloud platform itself is not distributed there. The older Botpress v12 is open source under AGPLv3 with a proprietary option. Chatbase is closed source and offers a public API v2 instead.
They meter differently. Chatbase Hobby is $40 a month for 700 message credits and Standard is $150 a month for 4,000 credits, with voice and the Helpdesk. Botpress Plus is $150 a month billed annually for 250 conversations and $25 of AI usage. A conversation can contain many messages, so compare against your own traffic: count monthly conversations and average replies per conversation before choosing.
Chatbase holds SOC 2 Type II and is GDPR compliant; HIPAA is available on the Enterprise plan only, under a BAA with zero data retention. Botpress shows a SOC 2 Type II badge on its Enterprise page and states compliance with EU data protection and US healthcare data privacy standards. For regulated data, request each vendor's report and BAA terms through sales before launch.
Yes, on the Standard plan and up. Chatbase agents can hold real-time voice conversations, answer inbound calls through Twilio numbers or your own SIP trunk, run outbound campaigns, and transfer a live call to a human.
Yes. Standard and higher plans include the built-in Chatbase Helpdesk, a shared inbox where human agents can take over in real time while the customer stays in the same chat. Chatbase can also create a ticket for follow-up or hand off live chat to Salesforce or Zendesk.
Often yes, if your workflow is centered on a third-party helpdesk, internal docs, ticket triage, or agent assist.
Use Fin when you want a customer-service AI agent inside a support suite and are comfortable with Intercom's operating model and pricing structure.
Chatbase. It is configuration-driven, so a marketer, founder or support lead can add content, set instructions and switch on actions without writing code. Botpress assumes engineering involvement and is better when you need custom logic.
Chatbase uses public subscriptions: Hobby at $40 a month, Standard at $150 and Pro at $500 on monthly billing, or $32, $120 and $400 a month on annual billing, with 700, 4,000 and 15,000 message credits a month. Credit use depends on the AI model, so a larger model gives fewer replies per credit. Fin charges per resolution (around $0.99 each, with minimums) and adds a per-seat helpdesk cost on Intercom's own suite. Fin can cost more as volume grows, but it ties spend to outcomes inside a support workflow.
eesel AI is built for helpdesks and knowledge bases. It connects to tools like Zendesk, Freshdesk, Intercom, Slack, Notion, Shopify, Salesforce, and HubSpot, and learns from your existing tickets and docs.
Use Apify to crawl and structure the source content, then choose the chatbot platform that fits your support workflow. See data for AI and RAG for the corpus-building side.