Chaindesk: the no-code support chatbot now known as Askli — Customer Support | AI Collection

Click the Chaindesk listing and you land on a site that calls itself Askli. Same domain, same product, new name. It is the third one this tool has worn: it started life in 2023 as Databerry, an open-source LLM agent builder from developer Georges Petrov, became Chaindesk, and now trades as Askli. If you came looking for Chaindesk, you are in the right place — just don't be thrown by the logo.

Underneath the name changes, the job has stayed the same: a no-code way to build a support chatbot trained on your own content. You point it at your docs, help center, and website, and it answers customer questions from that material instead of guessing. For a small team drowning in repetitive tickets, that is the whole pitch.

What it actually does

The setup follows four steps the homepage lays out plainly. You import your data, from files, Notion, Google Drive, a Zendesk help center, YouTube, or any public URL. You give the agent a persona and some goals so it sounds like your brand. You deploy it, either as an embeddable widget or inside tools you already use. Then you monitor the conversations and step in when the bot is out of its depth.

That last part matters more than it sounds. The platform includes human handoff, so an agent can take over a live chat when the AI hits something it shouldn't answer. Paired with what the company calls Trustworthy AI — safeguards that keep replies anchored to your knowledge base — the design is clearly aimed at reducing the two things that sink support bots: confident wrong answers and dead-end conversations with no escape hatch. The vendor claims it can resolve up to 80% of support queries; treat that as a marketing ceiling rather than a promise, but the grounding-first approach is a sensible way to chase it.

Answers are generated with GPT-4 and GPT-4o-mini, and there is an Auto-Sync option that re-ingests your sources so the bot doesn't drift out of date as your docs change. Multilingual support is built in, which opens it up to customers who don't speak English.

Channels it plugs into

A support bot is only useful where your customers already are, and this is one of the tool's stronger areas. Beyond the website widget, it connects to WhatsApp, Slack, Telegram, Messenger, Crisp, Zendesk, and Zapier, and there is a Shopify app for store owners. The integration list reads like the places an SMB support team actually lives, rather than a wish list of enterprise connectors nobody turns on.

The open-source lineage is worth weighing here too. Databerry was built on LangChain, Qdrant, OpenAI, and a standard React and Vercel stack, and the original repo is still public. That history is a small but real trust signal: this isn't a wrapper that appeared last quarter.

What Askli costs

Pricing is credit-based and starts at nothing.

The Free plan is $0 and gives you 200 message credits a month on gpt-4o-mini, one agent, one datastore, and a website loader capped at 25 pages. It is enough to prototype a bot and see whether the answers hold up, with one catch worth knowing: on the free tier, agents and datastores are deleted after 14 days of inactivity, so it is a trial sandbox, not a place to park a low-traffic bot.

From there, Growth is $49 a month (20,000 message credits, two agents, a 250-page loader, ten team seats, and your branding removed), Pro is $99 a month (40,000 credits, five agents, a 1,000-page loader, 25 seats), and Enterprise is $499 a month (400,000 credits and much higher limits). Annual billing knocks 20% off. The tiers scale on the levers that matter for a chatbot — message volume, number of agents, and how much of your site it can crawl — so picking a plan is mostly a question of how busy your support queue is.

One small note from the pricing page itself: a few of the storage figures on the higher tiers are obviously placeholder values (a "words storage" number in the tens of millions, a file-size cap in the millions of megabytes). It doesn't change what you pay, but it is the kind of unpolished detail that stands out when you are deciding whether to trust a vendor with your support channel.

Who it's for

This fits a solo founder or a small support team that wants a grounded, on-brand chatbot without hiring a developer or wiring up an API. The no-code flow, the free starting tier, and the spread of consumer-messaging integrations all point at that buyer. Teams with deep customization needs or unusual workflows will likely bump into the platform's edges and need more hands-on support, which independent write-ups flag as a real limitation.

Where it falls short

The honest considerations are less about missing features and more about polish and continuity. The name has changed twice in roughly two years, and while the product survived each rebrand intact, that churn is something to factor in if you care about long-term stability and consistent documentation. The pricing page's placeholder numbers reinforce that impression of a small team moving fast.

Beyond that, the usual caveats for this category apply and are echoed by third-party reviews: the quality of the bot depends heavily on the quality of the data you feed it, and there is real upfront work in curating that content and tuning the agent before it earns its keep. Independent buzz is also thin — the Show HN posts drew little discussion, and most of what turns up in a search is aggregator and directory coverage rather than deep user reports, so social proof is limited.

None of that is disqualifying. For the price, the integration coverage and grounding safeguards make it a credible pick in a crowded field. Overall: 4/5 for small teams and solo founders who want a no-code support bot that sticks to the facts, with the rebrand history and thin independent track record the main things holding it back from a clean recommendation.

Sources consulted

• Chaindesk / Askli homepage — what the product does, the four-step setup, models, channels, and the Askli rebrand
• Askli pricing — the Free/Growth/Pro/Enterprise tiers, credit limits, and the free-plan 14-day inactivity rule
• Databerry on GitHub — the open-source origin (2023), founder, and tech stack behind the tool
• Hacker News search: Chaindesk / Databerry — the Show HN posts and their limited traction
• Askli (Chaindesk) on AI Agent Store — independent pros, cons, and alternatives
• Chaindesk review on AIChief — a third-party review of use cases and limitations
• Askli on the Shopify App Store — confirmation of the Shopify deployment channel
• Wayback Machine snapshot — archived history confirming longevity