TextCortex
TextCortex is a Berlin-based platform for building AI agents grounded in company knowledge. Teams create agents without code, connect them to their documents and apps, and get answers that cite the files they came from. Each agent can run on a model you pick from the major providers, and you can call agents in chat, @mention them in Slack or Microsoft Teams, or schedule them to run on their own.
TextCortex started out as a writing assistant and browser extension, with a chat assistant called Zeno. Today it is sold to companies, with industry pages for manufacturing, consulting, M&A, private equity and legal teams. Anyone who used it years ago for paraphrasing and blog drafts will find a different product, and the alternatives below cover both groups.

An answer with numbered citations and the source PDF open beside it, from TextCortex's help center.
Building agents without code
You set up an agent in a visual builder. Give it a name, a background and rules, choose its default model, attach knowledge bases, and switch on capabilities such as web search, data analysis and image generation. A preview chat on the same screen shows how it responds before you save. Finished agents can be shared with colleagues or the whole company, and the platform also ships prebuilt agents for research, presentations and other common jobs.
Knowledge search and cited answers
Knowledge search connects company sources, including document stores and wikis, and combines them with web results for deeper research. Answers carry numbered citations, and clicking one opens the source document with the relevant passage highlighted. You can also drop in spreadsheets or reports and have an agent analyze, summarize or chart the data.
Automations and scheduled agents
Agents don't have to wait for a prompt. A node-based editor builds multi-step workflows, for example a daily research report that is written by one agent and emailed by another, with a Teams notification at the end. Scheduled agents run hourly, daily or weekly and send results to your inbox or as in-app notifications. Its browser extension brings the same agents to more than 30,000 apps and sites, and Zapier and Make connect thousands more.
Models, privacy and on-premise
The model hub covers current models from OpenAI, Anthropic, Google, Mistral and DeepSeek, so you aren't tied to one provider. TextCortex says it never uses customer data for training or fine-tuning. Services are hosted in the European Union, the company holds ISO 27001 and SOC 2 Type II certification, and larger customers can choose dedicated single-tenant infrastructure or run TextCortex on servers in their own data center with open-weight models. An optional add-on, Cortex Guard, checks prompts and files sent to ChatGPT, Claude and Gemini in the browser for personal data, secrets and prompt injection.
Plans and pricing
There's no free plan. A 14-day free trial lets you test the platform first, and paid plans are priced per user, billed monthly or yearly:
| Plan | What you get |
|---|---|
| Premium | Knowledge bases, custom agents, agent skills and connectors, all major models with included usage, up to 10 GB of file storage and workflow automations |
| Enterprise | Everything in Premium plus SSO, SCIM and SAML, dedicated deployment, centralized billing and priority support, with custom pricing |
Model usage on Premium has limits, and extra usage can be added. Accounts can't be shared, so each team member needs a seat.
Is TextCortex free?
Only for the first 14 days. After the free trial, each user needs a Premium seat, or the company needs an Enterprise contract.
Is TextCortex still an AI writing assistant?
Only in part. You can still use agents to draft emails, reports and presentations, but the product is now built around company knowledge, agents and automation rather than copywriting templates. For paraphrasing and grammar checks, QuillBot and Grammarly are closer to the old TextCortex.
Which AI models does TextCortex use?
It offers current GPT, Claude, Gemini, Mistral and DeepSeek models, and you can set a different default model for each agent. The on-premise version runs open-weight models such as DeepSeek, Qwen and Gemma on your own hardware.
Does TextCortex use my data to train AI?
TextCortex states that customer data is never used to train or fine-tune models, and that its services, including third-party models, are hosted on servers in the EU.
What are the best TextCortex alternatives?
For company-wide AI, the main options are ChatGPT, Claude and Microsoft Copilot, with Mistral AI for European or self-hosted setups. For agents that act across your apps, look at Lindy, Relevance AI and n8n. Individuals who want many models in one place often choose Poe or Sider AI.
Alternatives to TextCortex
Because TextCortex now targets companies, the best alternative depends on why you're looking. Start with your situation:
| Your situation | Alternatives to consider |
|---|---|
| Your company runs on Microsoft 365 | Microsoft Copilot |
| You want a widely used assistant with business plans | ChatGPT, Claude |
| You need a European provider or self-hosting | Mistral AI, n8n |
| Your team's knowledge already lives in Notion | Notion AI |
| You want agents that complete tasks across your apps | Lindy, Relevance AI |
| You're one person who wants many models in one subscription | Poe, Sider AI |
| You used TextCortex for rewriting and grammar | QuillBot, Grammarly |
For teams and companies
- ChatGPT: OpenAI's assistant for writing, research, files and data analysis, with projects, connected work apps and business plans for teams. It runs on OpenAI models only.
- Claude AI: Anthropic's assistant for long documents, writing, research and coding, with team plans and tools such as Cowork and Claude Code.
- Microsoft Copilot: Built into Word, Excel, PowerPoint, Outlook and Teams on paid Microsoft 365 plans, so it works where Microsoft-based companies already keep their files.
- Mistral AI: French AI company with the Le Chat assistant and open-weight models you can deploy yourself, a good fit for teams that care about data sovereignty.
- Notion AI: Agents inside the Notion workspace that write, update databases and manage projects using the docs and wikis your team already keeps there.
- Lindy: No-code platform where you describe a task in plain English, connect your apps and let an AI agent handle work such as email triage or lead research.
- Relevance AI: No-code multi-agent platform built for sales and go-to-market teams, with agents that qualify leads, research prospects and update the CRM.
- n8n: Workflow automation and AI agent platform with 500+ integrations, code steps and human approvals, available as a managed cloud or self-hosted.
- Adapt: Open-source platform for company agents with shared knowledge, approvals for sensitive actions and no per-seat pricing.
For individuals and writers
- Poe: One chat interface for hundreds of text, image, video and audio models, plus custom bots, without separate subscriptions for each provider.
- Sider AI: Browser sidebar with ChatGPT, Claude, Gemini and other models for chatting, summarizing pages and writing on any site, close to how the old TextCortex extension worked.
- QuillBot: Paraphrasing tool with several rewriting modes, plus a grammar checker, summarizer, plagiarism checker and citation generator.
- Grammarly: Grammar, clarity and tone suggestions as you type in the browser and desktop apps, with generative AI for drafting and rewriting.

TRY IT