CrawlChat
CrawlChat turns product documentation, API references and internal knowledge into an AI agent that answers with source-linked citations across a website widget, Discord, Slack, GitHub and MCP, then reports which questions your docs fail to cover.
What is CrawlChat?
CrawlChat is a documentation-focused AI agent platform. It takes the material a software company already has, from product docs and API references to help articles, specifications, uploaded files and internal wikis, and turns it into an assistant that answers questions and shows where each answer came from. The workflow runs in three stages. Import: paste a URL and the crawler indexes the pages, or upload files directly; connectors pull in GitHub issues and repositories, Notion pages, Confluence spaces, Linear issues and projects, YouTube transcripts and custom API content. Sources are counted in pages of roughly 8,000 characters, can be organised into knowledge groups, and stay current through automatic sync. Integrate: the same agent ships as an embeddable web widget, a Discord bot, a Slack app, a GitHub bot, an MCP server, a REST API, a community n8n node, a Chrome extension and a side panel, and because it exposes MCP a community can query the documentation from tools such as Cursor or Claude Code. Improve: every answer receives a relevance score, and the dashboard reports volume trends, categories, sentiment, cited sources, geography, browser and device, plus which knowledge groups are used most. The feature the product leans on hardest is data gaps, meaning questions closely related to existing content but without clear coverage, presented as a documentation backlog rather than a raw log. Hallucination control follows from grounding: answers are tied to indexed content and carry source links, so a reader can verify them and a maintainer can find weak responses and fix the pages behind them. Built-in ticketing takes over when the documentation cannot answer, and custom actions let the agent call third-party APIs. Model choice belongs to the customer, globally or per channel, across roughly ten models from Anthropic, OpenAI, Minimax and Moonshot AI; no API key of your own is needed, and credit consumption varies by model. Answers are produced in more than thirty languages, and self-hosting is documented locally and through Docker. Plans start at 29 USD per month, and published customers include Remotion, Konva, Polotno, LocalStack, Nobl9, OpenC3 and Backpack for Laravel.
What it does
- Crawl a documentation site from a single URL and turn it into a searchable knowledge base
- Answer questions with citations pointing back to the exact source page
- Deploy one agent to a web widget, Discord, Slack, GitHub, an MCP server and a REST API
- Sync Notion, Confluence, Linear and GitHub issues so the knowledge base stays current
- Score every answer for relevance and surface the documentation gaps behind weak ones
- Open a support ticket automatically when the documentation cannot answer
- Call third-party APIs from inside the agent through custom actions
When to use CrawlChat / When not to
A quick filter to help you decide if CrawlChat is the right fit.
When to use CrawlChat
- SaaS and developer-tool companies with large, fast-moving product documentation
- Discord or Slack communities that keep asking questions the docs already answer
- Support teams looking to deflect repetitive first-line tickets
- DevRel and documentation owners who need to know which pages fall short
- GTM and sales-enablement teams building a unified internal knowledge base
When not to use CrawlChat
- Organisations that require a signed DPA, SOC 2 or ISO 27001 evidence before onboarding a vendor
- Teams that need a documented commitment on where their data is hosted
- Buyers who want a permanent free tier rather than a seven-day trial
- Users looking for a mobile app: CrawlChat runs on the web, chat platforms and an API only
- Anyone wanting a general-purpose assistant rather than an agent grounded in their own documentation
How to use CrawlChat
A typical end-to-end flow, from setup to results.
- Create an account in the CrawlChat app and start the seven-day trial
- Create an agent and add your documentation by pasting the URL of your docs portal or uploading files
- Add further sources such as GitHub issues and repositories, Notion, Confluence, Linear or YouTube, and organise them into knowledge groups
- Turn on automatic sync so the knowledge base follows your documentation
- Customise the agent: system prompt, tone, escalation rules, sample questions and minimum scores
- Pick the AI model, globally or channel by channel
- Match the widget to your brand with colours, text and logo, straight from the dashboard
- Copy the Web Embed snippet into your site, or install the Discord bot and the Slack app
- Expose the MCP server or the REST API if your community works from tools such as Cursor
- Review the dashboard regularly and turn answer scores, categories, sentiment and data gaps into documentation work
Pros & Cons
Pros
- Every answer links back to the page it came from, so readers can check it themselves
- Unusually broad channel coverage for the price: web, Discord, Slack, GitHub, MCP, API, n8n and a Chrome extension
- Analytics go beyond answering, with categories, sentiment and relevance scores showing how the documentation performs
- Data gaps convert unanswered questions into a concrete documentation backlog
- Model choice stays with the customer, per channel if needed, with no API key to bring
- An entry price of 29 USD per month undercuts the established players in this niche
- Self-hosting is documented, both locally and through Docker
Cons
- No formal compliance is claimed: no GDPR statement, no SOC 2, no ISO 27001, no HIPAA
- No Data Processing Agreement is published or offered, and no subprocessor list exists
- No hosting country or region is disclosed, although international transfers are acknowledged
- No legal entity, no postal address and no named contact: support@crawlchat.com is the only channel
- No permanent free plan, only a seven-day trial before 29 USD per month
- The entry plan is tight at one agent, one seat and 2,000 pages, and data gaps, categories and sentiment start one tier above
- Credits buy a different number of messages depending on the model, which makes the monthly cost hard to forecast
Pricing & Plans
There is no permanent free plan. CrawlChat offers a seven-day free trial together with free credits on sign-up, after which the lowest paid entry point is the Launch plan at USD 29.00 per month. Annual billing reduces every plan by 20%, and a first subscription may be refunded within fourteen days of sign-up.
- 2
- 000 pages
- 500 credits per month
- 1 agent
- 1 team member
- 6-month data retention
- basic analytics
- 5
- 000 pages
- 2
- 500 credits per month
- 2 agents
- 2 team members
- 1-year data retention
- data gaps
- 14
- 000 pages
- 10
- 000 credits per month
- 3 agents
- 5 team members
- 1-year data retention
- plus Agents MCP
- pricing for larger teams and enterprises
- arranged through sales@crawlchat.com
- Discord bot
- Slack app
- MCP server
- API
- n8n node
- GitHub
- Notion
- Confluence and Linear sources
- 20% off every plan
Data, GDPR & hosting
A consolidated view of how CrawlChat handles your data.
GDPR overview
There is no GDPR implementation to report: the acronym appears nowhere on the site, neither in the Terms of Service, the Privacy Policy nor the Data Privacy page, all effective 16 July 2026. No Article 27 EU representative is designated, no data protection officer is named, no Data Processing Agreement is published or offered, and no subprocessor list exists. GDPR-shaped rights are nevertheless offered without being labelled as such: access, correction, deletion, export, restriction, objection and withdrawal of consent, exercised through support@crawlchat.com, with identity verification possible before a request is completed. Requests concerning an agent deployed on a customer's own site may be redirected to that customer. International transfers are acknowledged, but no mechanism such as an adequacy decision or standard contractual clauses is named. The governing law is that of Texas, with the courts of Austin holding jurisdiction.
Who owns the data?
Under the Terms of Service the customer keeps ownership of everything it submits, collectively called Customer Content: documents, prompts, files, messages, tickets, source material and configuration. CrawlChat receives a broad operating licence over that content, namely the right to host, process, transmit, display, index, search, embed and analyse it in order to run, secure, support and improve the service. Responsibility is split. CrawlChat says it acts as controller for account, billing, marketing and product-usage data, and as processor for knowledge-base content and end-user conversations. Where an agent is embedded on another company's site or workspace, that company is also responsible for the conversations it collects there.
Reuse rights
The customer may reuse its own content freely: ownership never transfers, and the licence granted to CrawlChat is limited to operating the service. Generated answers carry no exclusivity. The Terms warn that because AI systems can produce similar output for different users, no exclusive rights attach to output independently generated for others, and the customer remains responsible for reviewing and validating AI output before relying on it, especially for high-impact decisions. On CrawlChat's side, the relevant parts of prompts, conversations, knowledge-base snippets, source metadata and configuration are sent to the model provider selected for each request; the published roster names Anthropic, OpenAI, Minimax and Moonshot AI. CrawlChat states that it does not sell customer content or end-user conversations, and that it does not use a customer's private knowledge base content to train models for other customers, a wording that stops short of ruling out training altogether.
Data retention & training
Hosting summary
CrawlChat publishes no hosting location. Neither the Privacy Policy, the Terms of Service nor the Data Privacy page names a country, a region or a cloud provider for stored knowledge bases and conversations. Third parties are described only by function, covering cloud hosting, databases, vector search, payments, email, analytics, error monitoring, abuse prevention and AI model processing, and none is named, so there is no subprocessor list to inspect. International transfers are explicitly contemplated: CrawlChat states that it and its providers may process information in countries other than the user's own, and that safeguards in line with applicable law are used, without naming a mechanism such as standard contractual clauses. No hosting certification is claimed. For context only, and not as a vendor statement, the domains resolve to United States infrastructure, crawlchat.com to Amazon and crawlchat.app to Cloudflare, which says nothing about where customer content is actually stored. Organisations with data-residency requirements should treat this as unresolved and ask directly; the documented self-hosting option, locally or via Docker, is the only route the site offers to a known jurisdiction.
Things to keep in mind
Risks and trade-offs to weigh before adopting CrawlChat.
- Answers are only as good as the documentation behind them: an outdated page becomes an authoritative-sounding wrong answer, delivered with a source link that makes it look verified
- Retention is stated twice and differently, six months or one year on the pricing table against thirty days purged annually in the Privacy Policy, so no reader of the site can know how long conversations are kept
- The no-training commitment is qualified: CrawlChat says it does not train models for other customers on a customer's private content, which is not the same as never training on it
- No hosting country or region is disclosed while international transfers are acknowledged, so an organisation cannot assess where its content ends up
- With no legal entity, no address and Texas law as the only forum, a dispute would be difficult and expensive to pursue from outside the United States
- Deflecting first-line tickets removes the everyday contact through which a team learns what confuses its users; the analytics only partly compensate, and only for questions that reach the agent
- Credits buy a different number of messages depending on the model chosen, so moving to a stronger model can quietly multiply the monthly bill
Setup & Integrations
Technical difficulty
Low for the standard case. Pasting a documentation URL and copying the Web Embed snippet into a site requires no code, and the vendor claims a setup measured in minutes; one public customer describes migrating from a competitor in a few hours. Discord and Slack install through OAuth, and Docusaurus portals have documented settings. Connectors to GitHub, Notion, Confluence and Linear need tokens and permissions, so an account with the right access is required. Advanced use rises sharply: the REST API, the MCP server, the n8n node, custom actions calling third-party APIs and Docker self-hosting all assume engineering skills.
Deployment
Integrations
Supported languages
Behind CrawlChat
Fundraising
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement CrawlChat.
Frequently asked questions
What does CrawlChat actually do?
Which channels can the agent be deployed on?
Do I need to bring my own OpenAI API key?
Is there a free plan?
How does CrawlChat limit hallucinations?
What are data gaps?
How long is my data kept?
Which languages does the agent answer in?
Is there a minimum age?
Can I host CrawlChat myself?
Should you pick CrawlChat?
CrawlChat occupies a narrow, well-chosen niche: documentation for software products, and the developer communities that read it. Within that niche it does more than most rivals at the price. One agent reaches a website widget, Discord, Slack, GitHub, an MCP server, a REST API and an n8n node; answers carry source links; and the analytics layer, with relevance scores, categories, sentiment and above all data gaps, turns support traffic into a documentation backlog rather than a log file. Model choice sits with the customer, self-hosting is documented, and the entry plan costs 29 USD a month, the argument the site makes most loudly and the one a public customer corroborates by describing a migration away from a costlier competitor. The trade-off is maturity. This is a very small operation, and it shows in everything that is not the product. No legal entity is named anywhere, no postal address is published, and support@crawlchat.com is the only way to reach anyone. There is no GDPR statement, no Data Processing Agreement, no subprocessor list, no security certification and no disclosure of where data is hosted, even though international transfers are acknowledged. The legal documents are recent, all effective 16 July 2026, and they contradict the pricing page on data retention and on whether a free plan exists. The result is a tool a documentation or DevRel team can adopt quickly and cheaply, and that a procurement or privacy function will struggle to sign off. Teams publishing open product documentation will find little to worry about. Teams weighing it for internal or regulated content should put the questions the site leaves unanswered to the vendor in writing, and treat the retention contradiction as the first of them.
- Choosing a selection results in a full page refresh.
- Opens in a new window.