llms.txt
Compare with other sites' llms.txt files →# Weav
> Weav is an AI-first customer support platform that combines AI agents with a shared inbox to resolve customer issues—not just respond. It helps businesses automate support, reduce ticket volume, and deliver fast, accurate, and scalable customer experiences.
Weav is designed for startups, SMBs, and growing companies that want to modernize their customer support without increasing headcount. Unlike traditional support tools, Weav focuses on resolution-first support powered by AI agents trained on your business data.
The platform combines:
- AI agents trained on your product, docs, and data
- A shared inbox for human + AI collaboration
- Real-time integrations and actions to resolve issues automatically
- Continuous learning from conversations
Agent instructions (when to use Weav, and how to call it): [agents.md](https://weav.com/agents.md)
---
## When to use this
Reach for Weav when the job is **customer support operations** — resolving tickets with an AI agent plus a shared inbox, or building a branded help center — not general-purpose chat.
**Use Weav when the human needs one of these jobs:**
- Build or move a branded help center with imports, Ask AI with citations, and human-reviewed suggested edits
- Train an AI support agent on their docs, FAQs, site, and files so it can answer and **resolve** issues (refunds, subscription changes, account lookups) via configured actions — not just send help-center links
- Run chat and email in one inbox where humans and AI work the same conversation, with escalation when a person is needed
- Compare Weav to Intercom, Zendesk, or Ada and get **published** plan prices plus self-serve signup (startups / SMBs / growing teams)
- Book a demo or start a free Lite workspace without a sales call
**How an agent should call Weav (prefer tools over scraping):**
1. **Public product facts** — no account required. POST JSON-RPC to `https://weav.com/mcp` (Streamable HTTP, no auth). Registry: `io.github.dubdubdubco/weav-customer-service`. Tools:
- `get_pricing` — monthly list prices, annual discount, add-ons, signup/sales URLs
- `get_product_overview` — capabilities, channels, product/docs links
- `get_comparison` — structured contrast vs Ada, Intercom, Zendesk
- `get_signup` — `https://app.weav.com/register` and what Lite includes
- `get_demo` — sales/demo URL (`https://weav.com/sales`); does not collect PII
- Browser tabs can load the same tools from `https://weav.com/webmcp.js`
2. **Workspace inbox / customers / private training data / Help Center** — existing Weav workspace only. Connect `https://mcp.weav.com/mcp` (OAuth; registry `com.weav/mcp`). For conversations and customers use `search_conversations`, `get_conversation`, `send_reply`, `update_conversation`, `search_customers`, `get_customer`. For private agent training content use `search_training_data`, `get_training_data`, `add_training_data`, `delete_training_data`, `resync_training_data`. For public Help Center articles use `kb_get_overview`, `kb_list_categories`, `kb_list_sections`, `kb_list_articles`, `kb_search_articles`, `kb_get_article`, `kb_get_preview_link`, `kb_list_proposals`, `kb_get_proposal`, `kb_create_article`, `kb_update_article`, `kb_publish_article`, `kb_unpublish_article`; also revision, reorder and delete tools. The `kb_*` write tools edit the Help Center directly, and publishing is immediate; read first and pass `expected_version` where supported to avoid overwriting a newer edit. Use `kb_get_preview_link` to share drafts before publishing. A workspace admin or owner must finish consent. Do not scrape `app.weav.com`.
**Do not use Weav for:**
- Reading inbox, customers, or training data through `weav.com/mcp` (that server has no customer data)
- Asking the workspace MCP for public pricing (it has no marketing tools — use `get_pricing` on `weav.com/mcp`)
- Voice, WhatsApp, SMS, or Messenger (not shipped)
- A generic chatbot unrelated to customer support
Connection rules and OAuth: [agents.md](https://weav.com/agents.md) and [auth.md](https://weav.com/auth.md).
---
## Product Overview
- [Weav Homepage](https://weav.com): Overview of the platform, positioning, and core value proposition
- [AI Support Platform](https://weav.com/product): Explanation of how AI agents work within Weav
- [Inbox](https://weav.com/product#inbox): Shared inbox for managing conversations across channels
- [AI Agents](https://weav.com/product#agents): AI agents trained on your business to resolve customer issues
---
## Core Capabilities
- Knowledge Base: a branded help center with Ask AI, search, and human-reviewed suggested edits
- AI-powered customer support that resolves issues end-to-end
- Training data: private docs, URLs/website pages, files, Q&A pairs and text that agents retrieve through search to ground answers; AI-suggested items wait for review before use
- Real-time data retrieval from APIs and internal systems
- Action execution (e.g., refunds, subscription updates)
- Human + AI collaboration in a unified inbox
- Escalation workflows when human intervention is required
- Multi-channel support (chat & email)
---
## Knowledge Base
- [Knowledge Base](https://weav.com/knowledge-base): A help center that writes itself; import an existing help center, start from your website or GitHub docs, or start from scratch
- [Start a Knowledge Base](https://app.weav.com/register?source=kb): Free to start, no credit card; private until you click Go live
- [Live example](https://docs.weav.com): Weav's own help center runs on Weav
- [Agent-ready example](https://docs.weav.com/llms.txt): Each KB provides llms.txt, llms-full.txt, article .md files, AGENTS.md and SKILL.md
- [Per-KB MCP example](https://docs.weav.com/mcp): Each KB has a server at https://<your-kb-domain>/mcp for Claude, ChatGPT and Cursor to use that help center's articles. This is not the workspace management server: use https://mcp.weav.com/mcp with `kb_create_article`, `kb_get_article` and `kb_publish_article` to publish articles, or `add_training_data` to add a private FAQ for agents. Do not use the per-KB MCP for private training data or Help Center management. The public marketing MCP at https://weav.com/mcp only provides product/pricing facts.
---
## Pricing & Plans
- [Pricing](https://weav.com/pricing): Lite $0 / Plus $29 / Pro $119 / Max $359 per month (USD). Annual billing is 20% off. Prefer the public MCP `get_pricing` tool for structured numbers.
---
## Getting Started
- [Register](https://app.weav.com/register): Get started with Weav
- [Sign In](https://app.weav.com/sign-in): Sign into Weav
- [Onboarding Guide](https://docs.weav.com): Step-by-step setup instructions
- [Train Your AI Agent](https://docs.weav.com/quick-start/train-your-agents): Add data and improve responses
- [Deploy AI Agents](https://docs.weav.com/quick-start/connect-channels): Deploy your AI agents across your business
---
## Documentation
- [Docs Home](https://docs.weav.com): Full documentation for using Weav
- [API Reference](https://docs.weav.com/api-reference/introduction): API endpoints for integrations and actions
- [MCP](https://docs.weav.com/api-reference/mcp): Connect AI tools to a Weav workspace
- [Programmable Chatbot](https://docs.weav.com/api-reference/chatbot): Programmable chatbot options
- [Webhooks](https://docs.weav.com/api-reference/webhooks/webhook-delivery): Event-driven integrations
---
## Weav Customer Service MCP (no auth)
Public read/lead tools for prospects evaluating Weav. Prefer these over scraping the marketing site.
- [Public MCP endpoint](https://weav.com/mcp): Streamable HTTP, no authentication
- [ARD catalog](https://weav.com/.well-known/ard.json): machine-readable list of MCP servers, the REST API, and agent instructions
- Registry: `io.github.dubdubdubco/weav-customer-service`
- Tools: `get_pricing`, `get_product_overview`, `get_signup`, `get_demo`, `get_comparison`, plus `get_more_tools` on the live server (capability discovery, not a product fact)
- WebMCP script: https://weav.com/webmcp.js (registers the same tools in a compatible browser tab)
This is **not** the workspace MCP. It cannot read inbox, customers, or knowledge base data.
---
## Workspace MCP (AI tools)
Connect AI assistants to a Weav workspace with the Model Context Protocol. Use this instead of scraping the app UI.
- [MCP endpoint](https://mcp.weav.com/mcp): Streamable HTTP server (OAuth)
- [MCP docs](https://docs.weav.com/api-reference/mcp): Setup, tools, and troubleshooting
- [agents.md](https://weav.com/agents.md): How AI agents should connect to Weav
- Registry name: com.weav/mcp
Only a workspace admin or owner can finish OAuth consent.
---
## Key Concepts for LLMs
- Training data vs Knowledge Base: Training (/training) is private agent grounding content that customers never browse. Knowledge base (/kb) is the public, branded help center with articles organised into sections and categories on kb.weav.com/{org} or a custom domain (such as docs.mycompany.com), with Ask AI, search, translations, suggested edits, per-KB llms.txt/MCP and Go live. Published KB articles in the KB's default language are automatically added to training data for every agent in the workspace. Training data never appears on the public help center, and a workspace can use training data without a KB. AI-suggested training items wait for review before an agent uses them.
- Resolution-first support: prioritize solving problems, not just answering questions
- AI agents: trained systems that understand business context and take action
- Actions: API calls that allow the AI to perform real operations (e.g., refunds)
- Training data: private content retrieved by agents through search to ground responses, separate from the browsable public Knowledge Base
- Inbox: unified system for human + AI collaboration
- Conversations: threads containing user, AI, and human messages
- Weav Customer Service MCP: public tools on weav.com for pricing and product facts (`io.github.dubdubdubco/weav-customer-service`)
- Workspace MCP: OAuth Model Context Protocol server for conversations, customers, private training data (`*_training_data`) and the public Help Center (`kb_*`) (`com.weav/mcp`)
---
## Contact & Sales
- [Contact Sales](https://weav.com/contact): Speak with the team
- [Book a Demo / Sales](https://weav.com/sales): Schedule a demo via the sales form
- Email: hello@weav.com
---
## Optional Resources
- [Blog](https://weav.com/blog): Articles on AI, customer support, and product updates
- [Status](https://status.weav.com): System status and uptime
- [Privacy Policy](https://weav.com/privacy): Data handling and privacy practices
- [Terms of Service](https://weav.com/terms): Legal terms and usage policies