AI assistant trained on your business — answers questions about your company.
Business
Cody — the product at meetcody.ai, not to be confused with Sourcegraph's Cody code assistant — is a business-focused AI assistant that trains on your company's own documents, spreadsheets, wikis, and websites, then answers questions using that knowledge. Built by Meetcody Inc. (a subsidiary of Chatbot.com's parent company, Livechat Inc.), it entered the market in 2023 as one of the first "give an LLM your company data and let your team ask it questions" products and has stayed in that lane through 2026, adding polish rather than chasing scope.
The product's position in the 2026 market is unusual: it is not the deepest RAG platform (that title belongs to enterprise-tier products like Glean and Guru), not the cheapest (Notebook LM and NotionAI cover casual use for free-tier prices), and not the most technical (Sourcegraph, Vectara, and Weaviate all serve developers who want to build their own). Cody's actual customer is the small business owner or ops lead who wants to hand a shared inbox, a policy library, or a product spec repository to their team as a chatbot without hiring an engineer or negotiating an enterprise contract.
The one-line positioning: Cody is the internal-knowledge chatbot to pick when you are a small-to-mid business owner who needs a working "ask our company data" chatbot in an afternoon, not a six-month RAG platform migration. It is the product where the setup takes twenty minutes, the answers come from your uploaded PDFs and Google Docs, and the team can start using it the same day without training. If you are running a 50-person engineering org with a mature internal wiki, Cody is probably too shallow. If you are running a 12-person e-commerce brand with a Google Drive full of vendor SOPs, Cody is very likely the right size.
For AIQORA readers, Cody is the "second-order" AI product — not a general-purpose model like ChatGPT or Claude, but an application layered on top of them, sold on the promise that plumbing your business data into an LLM should not require a data engineering team.
Cody's feature surface is deliberately weighted toward "easy to set up, easy for a non-technical team to use." Enterprise depth is not the pitch.
Multi-source knowledge ingestion. Upload PDFs, DOCX, PPTX, Excel spreadsheets, plain text, and Markdown. Connect Google Drive, Dropbox, OneDrive, Notion, Zendesk, Confluence, and Intercom. Crawl a public or authenticated website. All sources feed into a single knowledge base your bots can query, and re-crawls run on a schedule you set. The connector set is broader than most "chatbot for your business" products and covers the shape of most small-business document sprawl.
Multiple bots per knowledge base. Create separate bots for separate audiences — a customer-facing product bot on your marketing site, an internal HR bot for policies, a sales enablement bot for the SDR team — each with its own personality, permissions, and scope of knowledge. The multi-bot model matches how small businesses actually deploy: not one universal chatbot, but a handful of specialists.
LLM choice. Cody proxies to multiple underlying models — GPT-4-class from OpenAI and Claude-tier from Anthropic across their current 2026 releases — with the model choice exposed as a setting. Users notice the difference on complex reasoning; Anthropic's models tend to answer more cautiously, OpenAI's tend to answer more expansively. Being able to pick without changing platforms is a real value-add.
Scoped knowledge and personality. Per-bot instructions tell each bot who it is, what tone to use, what topics to avoid, and what to do when it does not know an answer. Personality tuning is where a "generic chatbot" becomes "our team's chatbot" — Cody's implementation is straightforward and works.
Widget and embed options. Ship a bot as a floating chat widget on your marketing site, embed it in a page as an iframe, or connect it to Slack for internal use. The widget UI is fine, not spectacular — Intercom and Drift ship prettier UIs, but for internal tools nobody cares.
Source citations in answers. When Cody answers a question, it links to the source documents that informed the answer. This is the single most important RAG feature for building trust — users can click through to verify — and Cody's implementation is competent, if less polished than Glean's inline previews.
Access control and workspaces. Team members get seats, roles (owner, admin, editor, viewer), and can be scoped to specific bots. Not enterprise-grade IAM, but appropriate for the SMB target market.
Team analytics. Basic usage stats — questions asked, most-asked topics, unanswered questions — that help you understand what your team is actually using the bot for. The "unanswered questions" report is genuinely useful for identifying documentation gaps.
API access. REST API for integrating Cody bots into custom applications on higher tiers. Real API, not a marketing checkbox.
Zapier integration. For non-technical automation — pipe unanswered questions to a Slack channel, log conversations to a Google Sheet, escalate specific keywords to a human.
Cody publishes four public tiers. Pricing is per-workspace with seat and bot count as the main scaling axes.
| Plan | Price (USD/month) | Best for |
|---|---|---|
| Basic | $29 | Trying it out — 5 bots, 3 seats, 1,000 GPT-3.5-tier messages, 1GB storage |
| Premium | $99 | Growing teams — unlimited bots, 5 seats, GPT-4-tier messages included, 3GB storage |
| Advanced | $249 | Established teams — 10 seats, higher message quotas, priority support |
| Enterprise | Custom | 20+ seats, SSO, dedicated support, custom SLAs, API rate limits |
Cody's pricing is where the honest read gets uncomfortable. Basic at $29/month sounds fair until you realize the message quota caps at 1,000 mostly-cheaper-model messages — a small team of five will burn through it fast. Premium at $99/month is where the product starts to make sense: unlimited bots, GPT-4-tier answers, and a message quota that survives real usage. Advanced at $249/month is where SMBs with more than five seats or heavy usage land.
Compared to alternatives: Notebook LM is free (with Google-scale privacy tradeoffs), Guru's Starter tier starts around $15/user/month, Glean is enterprise-priced ($40-60/user/month with a $10k+ annual minimum for many deployments), and building your own RAG on Vectara or Weaviate is technically cheaper but costs engineering time. Cody's Premium at $99 total (not per user) is genuinely fair for the small-team job.
The honest tradeoff: Cody is priced for the buyer who values setup speed and support over per-user economics. A 30-person org will save money by looking at Guru or a custom Vectara setup; a 5-person org will save time by staying on Cody Premium.
Pros
Cons
Small e-commerce brands with vendor and SOP libraries. A 12-person e-commerce brand with a Google Drive full of vendor onboarding docs, product spec PDFs, and shipping SOPs. Point Cody at the Drive, deploy a bot to Slack, and the ops team stops interrupting each other for "how do we handle X" questions.
Agencies and consultancies with client documentation. Marketing agencies, consultancies, and service businesses with client-specific documentation get real value from per-client bots — one Cody workspace, one bot per client account, isolated knowledge scopes.
B2B SaaS companies with a public help center. A working "ask our docs" chatbot embedded in the marketing site or help center. Reduces support ticket volume for the common "where is this feature" questions. Cheaper than building it yourself, faster than migrating to Intercom's built-in bot.
Internal HR and policy chatbots. Employee handbook, benefits documentation, expense policies, IT setup guides — the SMB-scale internal knowledge that a 20-person company does not have HR budget to build a proper wiki for.
Real estate brokerages and franchise operations. Property databases, franchise operations manuals, agent training materials — Cody handles a mix of PDF, spreadsheet, and web content that most competitors do not ingest cleanly.
Course creators and content businesses. Turn a course library or a long-form content archive into a chatbot readers can query. Not a substitute for reading the course, but a real value-add.
Do not use Cody for: enterprise-scale internal search (Glean, Guru), developer-built custom RAG (Vectara, Weaviate, LlamaIndex), customer-facing sales chat (Intercom, Drift), or general-purpose company AI assistants where you want the assistant to also compose new content (ChatGPT Team, Claude with Projects).
Cody sits in a crowded category, and the honest read is that different alternatives win in different slices:
Notebook LM (Google) — free, powered by Gemini, generous file limits. Wins on price and integration with Google Docs. Loses on multi-bot deployment, widget embedding, and the "we own our data" enterprise conversation.
Guru — the polished veteran of the "verified answers" internal-knowledge category. Better on knowledge verification workflows, better on per-user pricing at scale, weaker on ingesting messy document repositories.
Glean — the enterprise-tier internal search product. Genuinely deeper semantic search, real IAM integration, priced accordingly. If you have a $10k+ annual budget for internal search, Glean is the sharper choice.
Intercom Fin — customer-facing AI chat trained on your help center. Better for support-desk deployment, priced per-resolution instead of per-message, worse for internal use cases.
ChatGPT Team with custom GPTs — the "just build a custom GPT" alternative. Cheaper if you have OpenAI accounts anyway, weaker on multi-source ingestion, no native widget embed, no separate bots per audience.
NotionAI Q&A — if your entire knowledge base already lives in Notion, NotionAI's Q&A across your workspace is the least-friction option. Weaker outside Notion.
Vectara / Weaviate / LlamaIndex + LangChain — the build-your-own-RAG path. Best per-unit economics at scale, real engineering cost to get right. Not a fair comparison to Cody unless you have RAG-experienced engineers.
For most AIQORA readers running a small business: Cody Premium at $99/month is the fast path. If you have a technical co-founder and time, Vectara + a light UI is cheaper long-term. If you have a $10k+ annual budget for internal knowledge tooling, Glean is the more polished tool.
Sign up at meetcody.ai — free trial available. The trial gives you access to the Basic tier feature set with limited messages. Enough to see if the product fits your team's shape.
Connect two knowledge sources. Start with a Google Drive folder and a website crawl, or a Notion workspace and a PDF library. Do not connect everything at once — a small, focused knowledge base outperforms a large messy one on answer quality.
Create one bot with a specific persona. Give it a name, a personality description, and instructions about tone and scope. "You are an internal support bot for the acme.co ops team. Answer questions about vendor SOPs and shipping policies. If you do not know, say so and suggest asking on Slack." That level of specificity substantially improves output.
Deploy the bot to Slack for internal use, or the widget for external use. Slack deployment takes about ten minutes. Widget embed is one line of JavaScript. Measure real usage for a week before optimizing.
Review the unanswered-questions report weekly. This is where Cody earns its ongoing value — the report tells you what your team is asking that your documentation does not cover. Filling those gaps improves answers immediately.
Skip the Basic tier for anything beyond initial evaluation. Premium is where the product starts to work.
Is Cody the same as Sourcegraph Cody? No — completely different products from different companies. Meetcody.ai is a business knowledge chatbot. Sourcegraph Cody is a code assistant for developers. Common naming confusion, worth double-checking which one you are evaluating.
Which LLM does Cody use? Cody proxies to multiple underlying models, with GPT-4-class and Claude-tier options exposed as settings on Premium and above. Basic tier defaults to a cheaper GPT-3.5-tier model.
Can I trust Cody's answers? Cody cites its sources in every answer, which lets users verify the reasoning. That said, RAG answers are only as good as the source documents — if your knowledge base has stale or contradictory content, so will the bot's answers.
How does Cody handle data privacy? Standard SaaS security posture — SOC 2 compliance, data encrypted in transit and at rest. Uploaded content is not used to train the underlying LLMs. Enterprise plans add SSO, custom retention policies, and DPA. For genuinely sensitive data (healthcare, legal), verify the current compliance certifications against your requirements.
Does Cody work in languages other than English? Multilingual documents ingest and query correctly, but Cody has no dedicated multilingual training tier. For Arabic-, Japanese-, or Chinese-first businesses, verify answer quality on real test queries before committing.
Can I use Cody for customer support? Yes — the widget deploys to your marketing site and can handle Tier 1 support questions. For high-volume support with escalation flows, Intercom Fin or Zendesk AI are more purpose-built.
Is there a free tier? No permanent free tier — Cody offers a free trial of the Basic tier, then $29/month minimum.
Cody on the Premium plan at $99/month is the sharpest internal knowledge chatbot for small businesses that need a working "ask our documents" bot in an afternoon. The multi-source ingestion, the multi-bot model, the source-cited answers, and the sub-day setup time add up to a product that fits the SMB job cleanly.
Where Cody stops being the sharpest answer: teams of 30+ where per-user pricing on Guru or Glean starts to make more sense; developer teams comfortable building their own RAG on Vectara or Weaviate for lower unit costs; and enterprise deployments with real IAM and compliance requirements Glean is built for. Cody is not competing there, and it should not.
The honest recommendation: start the free trial, connect one Google Drive folder and one website, deploy a Slack bot for your internal team, and measure a week of usage. If the "unanswered questions" report is telling you real things about your documentation and your team is genuinely using the bot instead of asking each other, commit to Premium annually. If usage is thin after a week, the fit is not there and your team is probably better served by ChatGPT Team with a custom GPT.
For the specific SMB job Cody was built for, it is the right size at the right price. Do not overthink it against enterprise tools you cannot afford, and do not overthink it against consumer tools that will not scale to your team.