Gallabox vs Manychat (2026)

Connectors cover CRM with HubSpot, Zoho, Salesforce, Pipedrive, LeadSquared, Kylas, Odoo and Sangam, commerce with Shopify, WooCommerce and Shopflo, payments with Razorpay, Stripe and Cashfree, logistics with Shiprocket and Shipway, and marketing with CleverTap, MoEngage, WebEngage, Fyno and Facebook Leads, along with Calendly, Zoho Books, Google Sheets, MiiTel, Pabbly, Zapier and generic webhooks.

They work in both directions: CRM push and pull nodes move records both ways, a Zoho lead panel renders CRM data inside the conversation, and payment integrations take money inside WhatsApp through catalog and checkout. On top sits the apiCall agent action, which issues HTTP requests of any standard method, from GET to DELETE, to any endpoint with operator-supplied headers and body and feeds the response back into the conversation. The agent invokes it when the condition the operator wrote is met, with a configurable fallback when the call fails, and actions chain in sequence.

Sourcegallabox.com docs Integrations overview, Connectors and Agent Actions pages; docs.gallabox.com/integrations/overviewread 2026-09-05

Flows write into the marketing stack through named connectors. Contacts collected in DMs go to HubSpot, Klaviyo, Mailchimp, ActiveCampaign, Kit and Flodesk, chat data logs to Google Sheets, and Hotmart sells courses inside Instagram and WhatsApp conversations. Zapier reaches more than 7,000 apps and Make builds scenarios across them.

The External Request action calls any HTTPS API with GET, POST, PUT or DELETE, custom headers and a JSON body, and maps the response into contact fields. Developers can also package an API as a Manychat Application with its own auth, actions and option sources. The AI reaches these systems through the flow it sits in, rather than choosing a tool itself in the middle of a conversation.

Sourcemanychat.com/integrationsread 2026-10-09

Two orchestration models run side by side. The rule-based Flow Builder is a canvas where nodes are arranged, a trigger set, and the flow tested and then published live. Its nodes come in four kinds, send, ask, action and flow-control nodes, and the flow-control nodes form the logic layer that branches conversations, adds delays, calls external APIs and routes into other flows through a Jump Flow node. Condition and branching depth is tiered, from basic through advanced to full custom.

The agent path orchestrates by instruction instead, chaining actions in sequence, so in one example the agent fetches an order through apiCall and then hands the conversation to a fulfillment team, with a fallback response configured for a failed call. Around both sit assignment rules, round-robin routing within teams, drip sequences with triggers, delays and conditional steps, and Smart Retry on broadcasts.

Sourcegallabox.com docs Flow Builder, Flow Control Nodes, Agent Actions and Sequences pages; docs.gallabox.com/ai-agents-and-bots/bots/flow-builderread 2026-09-05

Flow Builder chains triggers, messages, conditions, delays, actions and other flows on one canvas, with a Randomizer that splits contacts across up to 12 weighted paths, sequences that send over days, and rules that run actions on events.

An AI Step drops into any flow: the brand states a goal and background, the AI turns them into tasks, carries the conversation and saves answers into fields that later steps branch on.

On Instagram, up to three AI Goals follow an AI reply in a priority order the brand sets, and a lead capture ends by sending a message, starting another automation or assigning the conversation to a person.

Sourcehelp.manychat.com/hc/en-us/articles/14281187288860-manychat-ai-stepread 2026-10-09

Text channels are WhatsApp on the official Business API, Instagram DMs with public comment automation driven by configured intents, and a Web Chat widget that Gallabox controls outright, so no 24-hour messaging window applies. Voice covers human WhatsApp calling inbound and outbound with routing to the right agents, AI voice agents answering inbound WhatsApp calls, and AI voice on phone lines where the customer connects its own telephony provider.

Inbound triggers include Click-to-WhatsApp ads with conversion signals returned to Meta through the Conversions API, QR codes, chat links, a website click-to-chat button and the embedded widget. Outbound, broadcasts reach segmented audiences from ten thousand contacts a month to unlimited by tier, and drip sequences fire on a trigger and step through delays and conditions. Business hours, special days and holiday configuration bound when agents run, and agents carry per-channel greetings, instructions and availability.

Sourcegallabox.com docs Compare Channels, WhatsApp Calling, Broadcasts, Sequences and Comment Automation pages; docs.gallabox.com/concepts/channelsread 2026-09-05

Inbound social activity starts the work. Comments on posts and reels, DMs matching a keyword or a described intention, story replies and mentions, new followers, clicks on Instagram, Facebook and WhatsApp ads, Facebook Shop messages, QR codes and ref links each open an automation, and AI Replies and AI Comments answer as messages arrive, around the clock.

Sequences and broadcasts run on schedules, and outside apps start flows by posting to Manychat's webhook endpoint. Channels span Instagram, TikTok, Messenger, WhatsApp, Telegram, SMS and email. AI Replies, Comments and Goals run on Instagram only, the AI Step on Messenger, Instagram, WhatsApp, TikTok and Telegram, and intention recognition on all of those but TikTok.

Sourcehelp.manychat.com/hc/en-us/articles/23018283889180-manychat-ai-repliesread 2026-10-09

The Knowledge Base works as a RAG system, where the agent searches it when a visitor asks, retrieves the most relevant chunks, includes them in the prompt and answers from the customer's own documents. Vector storage runs on a managed Qdrant provider or against the customer's own OpenAI Assistant vector store.

Chunking is configurable, defaulting to 800 tokens with 400 tokens of overlap, and retrieval is configurable, defaulting to two documents per query at a relevance threshold of 0.7, with hybrid semantic plus keyword search and reranking enabled by default. Sources ingest as PDF, Word, CSV, Excel, text, Markdown and JSON, plus URL crawling. Documents show a Pending, Processing or Completed indexing state and can be re-indexed individually without rebuilding the base, and an agent can hold up to three Knowledge Bases. Knowledge base size is gated by tier, 10 MB on Essential and 100 MB on Advanced, with DOC and PDF on Essential widening to URL, PDF and DOC on Advanced.

Sourcegallabox.com docs Knowledge Base overview and add data sources pages; docs.gallabox.com/ai-agents-and-bots/chat-ai-agent/knowledge-base/overviewread 2026-09-05

A Knowledge library feeds AI Replies. Brands add text entries of up to 250,000 characters each and public web pages, scanned one page at a time, shown for review and editing, and active only once saved. Link entries can be refreshed when a page changes, entries can be updated or deleted at any time, and edits apply from the next conversation.

In the AI Playground every test answer lists the entries it drew on. Only visible page text is extracted, so images, video and PDFs are skipped, social profiles and Google Docs are refused, and file uploads are marked coming soon. The AI Step instead works from up to 10,000 characters of context written into the step.

Sourcehelp.manychat.com/hc/en-us/articles/25626595060124-manychat-ai-knowledgeread 2026-10-09

Within a session the Chat AI Agent holds full conversational context across multiple turns, whereas bot flows carry single-turn context and AI Reply a single message.

Outside the session, contact fields, conversation fields, tags and dynamic segments persist against the customer record indefinitely, and an escalated conversation hands a person the full history.

There is no memory product, and Gallabox does not say what an agent retains between separate conversations with the same contact, for how long, at what scope, or how an operator inspects or deletes it. The session is the unit throughout the analytics surface.

Sourcegallabox.com docs Chat AI Agent overview, Contacts and Segments pages; docs.gallabox.com/ai-agents-and-bots/chat-ai-agent/overviewread 2026-09-05

Inside a conversation the AI keeps the thread, adapting to what the follower says over several turns and waiting up to 10 hours for a reply before it stops. Answers it gathers, such as an email or a budget, are saved to system or custom contact fields that later flows and the Inbox read, so they belong to the contact record rather than to a memory the agent keeps.

AI Comments studies the account's past comment replies to match its tone, and AI Behavior holds the persona and guardrails as settings. Nothing carries what the agent learned about a person from one conversation to the next with its own scope, lifetime or deletion control.

Sourcehelp.manychat.com/hc/en-us/articles/14281227789468-power-up-your-chat-marketing-with-manychat-airead 2026-10-09

The operator writes the agent's Role, Goal and Instructions, stating topics to refuse, tone rules, escalation conditions and how to handle off-topic questions, so behavior is bounded by a policy the customer writes. Agent Actions limit what the agent can do, and Gallabox's answer to a wrong or harmful answer is to restrict the agent through Actions, which are individually gated and invoked only on conditions the operator specifies. AI Reply is a separate mode in which the AI drafts and a person reviews and sends, so every message can be approved before it goes out.

Escalation through assignTo and assignmentRule routes to a named user, team or bot, and the conversation lands in the Omnichannel Inbox carrying full history, captured customer details, the agent's stated reason for escalating and its notes. Availability hours and an agent active or inactive toggle bound when it runs at all. Web Chat has no Inbox for a live session, so assignTo and assignmentRule do not work there and the only escalation path is redirecting the visitor to WhatsApp.

Sourcegallabox.com docs Agent Actions, Chat AI Agent overview and AI Reply pages; docs.gallabox.com/ai-agents-and-bots/chat-ai-agent/agent-actionsread 2026-09-05

AI Behavior sets the AI's role, brand voice and explicit guardrails, the topics it should never touch and what it should never say or do, starting from a draft written off the account's bio and recent comments. Before AI Replies goes live, the brand tests it in the AI Playground and picks who takes conversations the AI cannot answer, and Assign to a person routes those to a team member or group.

Each skill has its own toggle, only Admins and Editors can change Manychat AI, and staff can pause automation for a contact from 30 minutes to indefinitely and resume it from the Inbox. AI Comments is tuned by approving or rejecting five sample replies. The product page says brands approve every AI comment reply, while the help center shows approval only at setup.

Sourcecommunity.manychat.com/product-updates/your-ai-now-speaks-your-language-meet-ai-behavior-9491read 2026-10-09

Customers control who inside their organization can do what, with multi-factor authentication enforced for all users, role-based access control that is team-based and channel-based, and custom roles carrying per-module permissions such as the developer permission governing webhooks and API keys, alongside phone and number masking, IP whitelisting restricting workspace access, and SSO and SAML on the Enterprise tier.

Security controls also include AES-256 at rest and TLS 1.2 or 1.3 in transit, a cloud web application firewall, SIEM logging, endpoint detection, periodic penetration testing, and PII-permission masking on webhook payload views. The pricing page says Gallabox is SOC 2 Type II compliant and the footer carries an AICPA SOC mark, while the documentation security page says the company is in the process of becoming SOC 2 and GDPR compliant, and no trust center, report request path or auditor letter is published.

Sourcegallabox.com docs Data Security, Two Factor Authentication and Users, Teams and Roles pages, gallabox.com pricing and footer; docs.gallabox.com/privacy-and-security/data-securityread 2026-09-05

Accounts have five roles with set permissions: Owner, Admin, Editor, Inbox Agent and Viewer. Inbox Agents work conversations, tags and contact fields without touching automations, Viewers see statistics read only, invitation links are single use and expire in 24 hours, and only Admins and Editors reach Manychat AI.

Manychat holds ISO/IEC 27001:2022 certification from BSI, a SOC 2 Type II report from A-LIGN available on request, ISO/IEC 42001:2023 for AI management, a CSA STAR listing and PCI DSS compliance under SAQ A for its payment integrations. Data is encrypted in transit with TLS 1.2 and 1.3 and at rest to FIPS 140-2, penetration tests run yearly and a private bug bounty runs on Bugcrowd. Single sign on, SCIM and two factor sign in do not appear.

Sourcemanychat.com/securityread 2026-10-09

Every agent session appears as a row in a sessions table, and opening one shows a detail drawer with the full Chat Log, the session ID, channel, contact, created date, session status with a drop reason where it applies, the AI credits the run consumed and the model that answered, so a specific reply can be attributed to a specific model and cost after the fact.

Sessions export to Excel with an export history, and a Conversation.AI.Analyze webhook pushes completed analyses into a system the customer already runs. Underneath sit an Activity Log of data interactions, audit trails the security page calls exhaustive, and conversation, broadcast, sequence, bot and template analytics. The same Analyze tab carries outcome evaluation and a Successful sessions metric, and the Agent Analyze surface sits on the Advanced tier.

Sourcegallabox.com docs AI Agent Analytics, Webhooks and Data Security pages; docs.gallabox.com/ai-agents-and-bots/chat-ai-agent/ai-agent-analyticsread 2026-09-05

Every message the AI sends sits in the contact's Inbox conversation beside the automations that ran, and an Overview tab in Manychat AI reports on AI performance. In the AI Playground each test answer shows the Knowledge entries behind it, and testers rate answers up or down with a reason. Inbox Analytics tracks response times, team efficiency and conversation volume. Live conversations do not carry that source view, and there is no log of agent decisions, tool calls or configuration changes for tracing a single reply after the fact.

Sourcecommunity.manychat.com/product-updates/better-faster-stronger-introducing-new-updates-to-ai-playground-9311read 2026-10-09

Gallabox's subprocessor table names seventeen processors and places every one in the United States, including AWS for storage and CDN, MongoDB for the database, Vercel and Heroku for hosting, and Redis and ClickHouse for infrastructure, and backups span multiple zones with a primary data center in the US.

For a vendor headquartered in India selling into Indian and Gulf support teams, that means the data leaves the region, and customers cannot choose where it sits. The Enterprise tier lists on-prem or private cloud and data residency as a pricing-page bullet, naming no region, mechanism or configuration surface, and the documentation describes neither.

Sourcegallabox.com docs Subprocessors and Data Security pages, gallabox.com pricing; docs.gallabox.com/privacy-and-security/subprocessorsread 2026-09-05

Manychat runs as a hosted service only, and all production data sits in AWS data centers in Frankfurt, with no other region on offer and no private cloud or on premises install.

AI features run on generative AI providers OpenAI, Anthropic, Google and Microsoft, all listed as processing in the US, and transfers from the EU, UK and Switzerland rest on the Data Privacy Framework and standard contractual clauses.

Full daily database backups are kept apart from the main data center, accounts inactive for 18 months are flagged for deletion, and personal data leaves active systems when an account is deleted.

Sourcemanychat.com/securityread 2026-10-09

The chatbot template library can be filtered by eight industries (ecommerce, healthcare, automobile, education, travel, real estate, finance and a popular set) and by seven use cases covering lead generation, booking and scheduling, sales, support FAQ, product recommendation, surveys and discovery.

Each template has its own URL, description, live Try action and usage count, among them a Lead Validation Bot, an emergency ambulance booking bot, a medical records retrieval bot, a membership booking bot and a vehicle rental recommendation bot.

Inside the product, the Chat AI Agent creation wizard asks for an industry and then a use case such as Lead Qualification, Sales, FAQ or Receptionist, and generates the agent's role, goal and instructions from that starting point.

Sourcegallabox.com chatbot templates library and docs Chat AI Agent overviewread 2026-09-05

Quick Automations are ready made setups a brand picks and fills in step by step, among them auto DM links from comments, generate leads with stories, grow followers from comments and a delayed welcome to new followers. Templates install whole automations, fields and tags from a link, creators and agencies build and protect their own, and Pro templates need a paid plan.

Manychat AI ships as separate skills: AI Replies answers questions, AI Comments answers positive comments, and AI Goals extends either by sharing a link, growing followers or capturing leads. The Flow Builder assistant drafts a whole automation from a plain description of the goal.

Sourcehelp.manychat.com/hc/en-us/articles/16654065283100-quick-automation-auto-dm-links-from-commentsread 2026-10-09

Each Chat AI Agent has an AI Model Settings section whose Model field lists the models enabled for the account's AI wallet, with the catalog naming the GPT-4o, 4.1 and 5 family, grok-4.3, Gemini 2.5 Pro and Flash, DeepSeek, Kimi and Qwen. Alongside the model the operator sets LLM Temperature from 0 to 2, defaulting to 0.1, and a Thinking Level of None, Low, Medium or High on reasoning-capable models.

Two provider versions exist, v2 defaulting to grok-4.3 and legacy v1 defaulting to gpt-4o-mini, with a one-click migration between them. Credit consumption varies by model at its own per-token rate, and the model used is recorded on each agent session. The v2 agent generation carries this selector, and v1 is deprecated.

Sourcegallabox.com docs Chat AI Agent overview, AI Model Settings and v1-to-v2 migration pages; docs.gallabox.com/ai-agents-and-bots/chat-ai-agent/overviewread 2026-09-05

The model behind Manychat AI is not named, and brands cannot pick one for AI Replies, Comments, Goals or the AI Step. The subprocessor list names OpenAI, Anthropic, Google and Microsoft for generative AI, and the AI terms cite OpenAI's policies and the Azure OpenAI code of conduct. Separately, flows can call ChatGPT, Claude or DeepSeek on the brand's own API key, choosing the model, prompt, temperature and token limit and saving the answer to a field. Those calls are single generations from the latest message, billed by the provider, and sit outside the Manychat AI skills.

Sourcemanychat.com/legal/service-providersread 2026-10-09

The API reference on Gallabox's api-docs subdomain is a Postman collection with request samples in more than twenty languages including cURL, Python, Node, Java, PHP, Go and Ruby, and the API is what a customer calls to query data or take actions in Gallabox, while webhooks push outward. Message and Contact APIs come with Basic and all APIs with higher tiers, with messaging rate limits of 100, 200 and 1,000 messages per minute by tier. Credentials have their own permission, held by custom roles that can view and manage webhooks and API keys.

Outbound, more than twenty-five subscribable events across contact, conversation, message, template and broadcast each carry the event name in an x-event-name header, with a configurable HTTP method, custom headers, an optional shared secret producing a base64 HMAC-SHA256 signature in x-gallabox-signature, automatic retry up to five attempts, and a delivery log holding status and payload for seven days.

Sourcegallabox.com docs Webhooks and pricing pages, api-docs.gallabox.com; docs.gallabox.com/settings/webhooksread 2026-09-05

A public REST API at api.manychat.com, with a Swagger console and Bearer token auth, covers the account (tags, custom and bot fields, flows, growth tools), sending (content and whole flows) and contacts (create, update, find, tag and set fields). Reads run at up to 100 requests a second, sends at 25 and writes at 10, and an account over the limit may be cut off for 24 hours.

Keys are generated under Settings on paid plans. The Dev Program adds Manychat Applications, JSON defined apps with an App Key per install, actions, option sources and triggers that outside systems fire through a webhook, with public apps listed at apps.manychat.com after review. Manychat publishes no MCP server of its own.

Sourcehelp.manychat.com/hc/en-us/articles/14959510331420-how-to-generate-a-token-for-the-manychat-api-and-where-to-get-parametersread 2026-10-09

The Test Widget is the step before going live, and operators run it before launch and again after configuring actions or connecting a Knowledge Base, while bot flows are tested in the Flow Builder and then published. Outcome evaluations, a named plan feature, score a conversation against the business outcome the operator defined, appear per session in the Analyze tab as a goal outcome, and roll up into a Successful sessions metric that measures whether the agent achieved its intended outcome rather than merely reaching an end state.

Dropped sessions carry a drop reason, and operators read them to find where a flow loses people, change the flow and republish it. Each evaluation deducts AI credits when a conversation is scored against an outcome. Outcome evaluations sit on the Advanced tier and custom evaluation rules on Enterprise, and there is no regression suite, versioned test set or prerelease scoring run against a fixed corpus.

Sourcegallabox.com docs AI Agent Analytics, Chat AI Agent and Flow Builder pages, gallabox.com pricing; docs.gallabox.com/ai-agents-and-bots/chat-ai-agent/ai-agent-analyticsread 2026-09-05

The AI Playground lets a brand chat with its AI as a follower would before going live, with suggested or custom questions. It flags questions the Knowledge library cannot answer and offers to add the missing information on the spot, shows which entries each answer drew on, and takes a thumbs up or down with a reason. Brands retest after each change, then go live.

Flows can be previewed before publishing, and the Randomizer splits contacts across weighted paths for A/B tests. There is no saved test set, scoring run or comparison between AI versions, and the Randomizer reports no results by path, so a change is judged by reading test conversations.

Sourcecommunity.manychat.com/product-updates/ai-playground-test-your-ai-detect-knowledge-gaps-and-go-live-with-confidence-9189read 2026-10-09

The route out of the platform is entirely programmatic. Named connectors write into CRM, commerce, payment and logistics systems, the apiCall action issues HTTP requests to any endpoint, bot flows carry an API node and a Google Sheets node, and outbound webhooks push events to systems the customer runs.

The closest capability is codeExecution, which is limited to a sandbox for basic mathematics and string manipulation, not full programming. Gallabox's customer-facing surface is a chat widget it renders, and nothing in the product controls a browser, runs a hosted session or automates an interface.

Sourcegallabox.com docs Agent Actions, Connectors and Flow Control Nodes pages; docs.gallabox.com/ai-agents-and-bots/chat-ai-agent/agent-actionsread 2026-09-05

Manychat works through the official messaging platforms as a Meta Business Partner and TikTok Marketing Partner, so DMs, comments and messages travel through the Instagram, Messenger, WhatsApp, TikTok and Telegram channels. No browser, desktop or computer is driven by the AI. Public web pages enter Knowledge as extracted text, a fetch for answers rather than an agent operating a site.

Sourcemanychat.comread 2026-10-09