Build AI Agents with GPT-OSS 20B
GPT-OSS 20B is the lightweight variant of OpenAI's open-source family—20 billion parameters with a 128K context window at just $0.07 per million input tokens. Fast inference, transparent weights, and no vendor lock-in make it ideal for high-volume support and cost-sensitive deployments.
7-day free trial · Cancel anytime · No commitment
Strengths
Also available
Open-source speed
at the lowest cost
Fast, transparent AI for teams that need throughput without premium pricing.
Fast inference
20B parameters deliver quick responses for real-time chat and support.
Budget-friendly
$0.07/M input—one of the lowest per-token costs available.
Open weights
Fully inspectable architecture with no proprietary lock-in.
Grounded answers
Responds from your knowledge base, not general training data alone.
Start building with GPT-OSS 20B today
7-day free trial · Cancel anytime · No commitment
Scale without
scaling costs
Handle high volumes with predictable spend and transparent AI under the hood.
Predictable pricing
Low, flat per-token costs that stay stable at any conversation volume.
Full agent support
Works with all InsertChat tools, embeds, and deployment options.
Reasoning capability
Built-in reasoning support despite its compact 20B size.
Multi-channel deploy
Embed on your site, use via API, or run in workspace.
Go from knowledge to
a live agent in minutes
Configure your agent
Pick a model, use prompt templates, and enable tools.
Deploy to channels
Launch a widget, embed in your app, or use the API.
Start with one agent and expand across teams, channels, and workflows.
What you get
with GPT-OSS 20B
Outcome-focused benefits you can measure in support, sales, and operations.
- Faster first responses without sacrificing grounded accuracy
- Lower per-conversation cost with a model built for throughput
- Reliable at high volumes—consistent quality from message 1 to 100K
- Scales from 100 to 100,000 conversations with predictable spend
What our users say
We deployed AI support in 20 minutes. Our response time dropped by 80%. Customers love it.
Marcus Weber
Head of Support, Notion
Finally, one place for all my AI needs. The ability to switch models mid-conversation is game-changing.
Sarah Chen
Product Designer, Figma
The white-label option let us offer AI services to our clients overnight. Revenue grew 40% in Q1.
Elena Rodriguez
Agency Founder, Digitale Studio
GPT-OSS 20B is included on every plan — pick the one that fits your team.
Frequently asked questions
Tap any question to see how InsertChat would respond.
Contact supportInsertChat
AI Support
Hey! 👋 Browsing Product questions. Tap any to get instant answers.
What is InsertChat?
An AI agent workspace that lets you build agents grounded in your knowledge and deploy them to web, app, or API. Connect tools and integrations to complete workflows.
What's the difference between an agent and an InsertChat agent?
A basic agent is prompt-only. InsertChat agents are grounded in your sources, configurable per use case, and able to use tools and integrations.
How do agents stay accurate and avoid hallucinations?
Ground your agent in a knowledge base your team controls and keep it fresh. Use analytics to find gaps and improve coverage over time.
What can I connect as knowledge?
URLs, sitemaps, documents (PDF and office files), media like YouTube and audio, and structured data. The goal is a clear source of truth for answers.
Do sources stay up to date?
Yes. Refresh sources on demand or set up scheduled refresh depending on the source type.
Can I control how the agent behaves?
Yes. Control prompts, model choice, tool access, and agent experience so behavior stays consistent.
Which AI models can I use?
GPT-5.2, Claude Sonnet 4.5, Gemini 3.0, Llama 4, Grok 4.1, DeepSeek V3.2, and more. Choose the model per chat, or use BYOK to manage provider access yourself.
Can I pick different models for different workflows?
Yes. Use a faster model for common questions and a stronger model for complex reasoning. InsertChat supports that balance per conversation.
Where can I deploy an agent?
Website widget, in-app embed, or API. Keep one agent setup and reuse it across channels.
Do I need coding skills?
No. Build and deploy AI agents using our visual builder. The embed code is one line of JavaScript.
Can I customize the branding and UI?
Yes. Customize the widget to match your brand. White-label options are available for a fully branded experience.
Can I use my own domain?
Yes. Custom domains are supported, typically via enterprise options.
Does InsertChat support voice?
Yes. Voice dictation and text-to-speech let users speak instead of type.
Does InsertChat support vision?
Yes. Enable vision for agents when images help clarify a request or context.
What tools and integrations are supported?
Zendesk, HubSpot, Shopify, WooCommerce, calendar booking, web search, Perplexity, and webhooks for your own systems.
Can I control which tools the agent is allowed to use?
Yes. Tool access is controlled per agent so you enable only what you need.
Can the agent hand off to a human?
Yes. Configure human handoff so the agent escalates when needed. Full conversation history is passed along.
Do you provide analytics?
Yes. Track chats, leads, feedback, and credits used. Find gaps in coverage and prioritize fixes.
Is it mobile friendly?
Yes. The widget and embeds work well on desktop and mobile with no separate experience needed.
What's the fastest path to a successful deployment?
Start with one agent and a small set of high-value sources. Iterate using real questions from analytics.
What is the fastest way to get started?
Create an account, upload one document, and ask your first question. Most teams go live in under 5 minutes.
Product FAQ
What is InsertChat?
An AI agent workspace that lets you build agents grounded in your knowledge and deploy them to web, app, or API. Connect tools and integrations to complete workflows.
What's the difference between an agent and an InsertChat agent?
A basic agent is prompt-only. InsertChat agents are grounded in your sources, configurable per use case, and able to use tools and integrations.
How do agents stay accurate and avoid hallucinations?
Ground your agent in a knowledge base your team controls and keep it fresh. Use analytics to find gaps and improve coverage over time.
What can I connect as knowledge?
URLs, sitemaps, documents (PDF and office files), media like YouTube and audio, and structured data. The goal is a clear source of truth for answers.
Do sources stay up to date?
Yes. Refresh sources on demand or set up scheduled refresh depending on the source type.
Can I control how the agent behaves?
Yes. Control prompts, model choice, tool access, and agent experience so behavior stays consistent.
Which AI models can I use?
GPT-5.2, Claude Sonnet 4.5, Gemini 3.0, Llama 4, Grok 4.1, DeepSeek V3.2, and more. Choose the model per chat, or use BYOK to manage provider access yourself.
Can I pick different models for different workflows?
Yes. Use a faster model for common questions and a stronger model for complex reasoning. InsertChat supports that balance per conversation.
Where can I deploy an agent?
Website widget, in-app embed, or API. Keep one agent setup and reuse it across channels.
Do I need coding skills?
No. Build and deploy AI agents using our visual builder. The embed code is one line of JavaScript.
Can I customize the branding and UI?
Yes. Customize the widget to match your brand. White-label options are available for a fully branded experience.
Can I use my own domain?
Yes. Custom domains are supported, typically via enterprise options.
Does InsertChat support voice?
Yes. Voice dictation and text-to-speech let users speak instead of type.
Does InsertChat support vision?
Yes. Enable vision for agents when images help clarify a request or context.
What tools and integrations are supported?
Zendesk, HubSpot, Shopify, WooCommerce, calendar booking, web search, Perplexity, and webhooks for your own systems.
Can I control which tools the agent is allowed to use?
Yes. Tool access is controlled per agent so you enable only what you need.
Can the agent hand off to a human?
Yes. Configure human handoff so the agent escalates when needed. Full conversation history is passed along.
Do you provide analytics?
Yes. Track chats, leads, feedback, and credits used. Find gaps in coverage and prioritize fixes.
Is it mobile friendly?
Yes. The widget and embeds work well on desktop and mobile with no separate experience needed.
What's the fastest path to a successful deployment?
Start with one agent and a small set of high-value sources. Iterate using real questions from analytics.
What is the fastest way to get started?
Create an account, upload one document, and ask your first question. Most teams go live in under 5 minutes.
Pricing FAQ
What do I pay for with InsertChat?
Pricing is based on how many agents you run, what knowledge sources you connect, and how much conversation usage you drive. Check the pricing page for current tiers.
How much does InsertChat cost?
Plans start at $29/month. Verify the latest pricing and included limits on the pricing page.
Is pricing per seat or per teammate?
Pricing is oriented around agents, sources, and usage rather than seats. Enterprise plans are available for larger teams.
Can I start small and upgrade later?
Yes. Start with self-serve, validate your use case, then scale up as needed.
Can I cancel anytime?
Yes. Cancel anytime with no long-term contract. Your data remains available for 30 days after cancellation.
Do you offer enterprise pricing?
Yes. Enterprise plans cover larger orgs, advanced requirements, and custom deployment needs.
Do you support annual billing?
Yes. Toggle to annual billing on the pricing page and save 20%. For invoicing or procurement workflows, contact us.
What happens if we hit limits?
You will get a notification before you hit a limit. Upgrade your plan anytime with one click, or reduce usage. Nothing stops working without warning.
What counts as a source?
A source is any connected item your agent learns from: a URL, document, YouTube link, or other knowledge input. Your plan determines how many you can connect.
Is there a max file size for uploads?
Yes. Plans enforce a max file size per upload. Check your plan details on the pricing page.
What are credits?
Credits budget your usage across conversations, sources, and tools. They keep costs predictable as you scale.
Do you charge for all messages, even unresolved ones?
Ground answers in your knowledge and enable tools only when needed to control cost. Check the pricing page for how usage is measured.
Can I control which models we use so costs do not spike?
Yes. Choose the model per chat to balance quality, speed, and budget for different workflows.
Do I need to buy separate subscriptions for each model?
No. Keep your agent workflow in one place and switch between GPT-5.2, Claude Sonnet 4.5, Gemini 3.0, and more without separate subscriptions.
What is BYOK?
Bring Your Own Key. Use your own provider API key for model access to consolidate billing or apply your own setup.
If I use BYOK, do I still pay InsertChat?
Yes. InsertChat still provides the agent product: knowledge, deployments, tools, analytics, and governance. Model provider usage is billed separately.
Can I use my own logo and domain on the $29 plan?
Yes. The $29 plan includes your own logo and custom domain.
Does InsertChat replace my support team?
No. Deploy agents for common questions and use handoff to humans for edge cases and high-stakes requests.
What's the ROI case for pricing?
Teams see value when agents handle repeat questions, capture leads, and complete actions via tools. Best ROI comes from grounding on your sources and iterating with analytics.
Can I test before I commit?
Start self-serve to validate your workflow. For guided proof-of-concept or enterprise requirements, contact us.
Do you have discounts for startups or nonprofits?
If pricing is a blocker, contact us with your context.
What's included in enterprise that self-serve might not cover?
Custom domains, BYOK, S3-compatible source sync, role-based access controls, and advanced integrations.
Can I use InsertChat on multiple websites or products?
Yes. Deploy agents across website, in-app embed, and API. See pricing details for how that maps to your plan.
If I stop paying, do I lose my data?
Retention depends on your plan terms. Contact us for specific export or retention guarantees before purchase.
What if InsertChat isn't a fit after we try it?
Share what didn't work and we'll help you decide whether to tune the setup or move on.
How do I start?
Sign up for a 7-day free trial with full access. Pick your plan after you see it working with your own content.
Security FAQ
Where is my data stored?
European servers. GDPR compliant, never used for training, and deletable at any time.
What is InsertChat, from a privacy standpoint?
An AI agent workspace that connects to your approved sources and uses model providers to generate responses. Privacy depends on what you ingest, what tools you enable, and who can access the agent.
What gets sent to AI model providers?
Your prompt and relevant context excerpts from connected sources are sent to the selected model provider to generate an answer.
Do you use our data to train models?
No. InsertChat never uses your data to train models.
Is my data isolated from other customers?
Yes. Data is scoped to your workspace and agents. Sources and conversations remain isolated.
Can I delete data?
Yes. Delete sources, conversation history, leads, and feedback at any time.
What data does InsertChat store?
Agent configuration, connected knowledge sources, and conversation data needed for the experience and analytics.
Can I keep an agent private?
Yes. Choose public or private agents depending on whether anyone or only authenticated users can access the embed.
Do you have role-based access controls?
Yes. Control who can manage agents and data with role-based access.
What is BYOK from a privacy standpoint?
You provide your own API key for model access. Prompts and context are still sent to the provider, so review their policies for your compliance needs.
Can I restrict what the agent can do?
Yes. Control tool enablement per agent to limit actions to only what is necessary.
Can we limit exposure of sensitive data in the agent?
Yes. Scope sources to what should be answerable and limit tool enablement to only what is required.
Can I export or audit what users asked?
Yes. Analytics show what people ask. Contact us for exports for audits or internal reporting.
Do you support GDPR?
Yes. Full GDPR compliance with Data Processing Addendum (DPA) available on request.
Can you provide a DPA?
Yes. Our DPA covers processing obligations, subprocessors, and deletion/return terms. Contact us to request it.
Do you list subprocessors?
Yes. Subprocessors are documented in the DPA. Request it or contact us for details.
How do you handle security questionnaires?
Contact us and we provide the right documentation for your team's review process.
Is InsertChat safe to embed on a public website?
Yes, when configured correctly. Ground answers in approved sources and keep tool access controlled.
What if we need a private deployment or special requirements?
Enterprise plans cover custom deployment, advanced controls, and procurement constraints. Contact us to discuss.
How do privacy requests work (access, deletion)?
Submit privacy requests through the contact page. We route them to the right process.
Where can I request security documentation?
Submit security inquiries through the contact page.
Do you support self-hosting?
Yes. Enterprise plans include self-hosting and bring-your-own-LLM options.
How do I evaluate InsertChat?
Start a free trial with non-sensitive data. When ready, request our security questionnaire and DPA.
Integrations FAQ
What integrations are available?
600+ integrations including Slack, Notion, Google Workspace, Salesforce, HubSpot, Zendesk, Shopify, WooCommerce, and Zapier. Our REST API allows custom integrations with any system.
Can I connect to Slack?
Yes. Deploy your agent directly to Slack so your team can interact in channels or DMs.
Do you integrate with HubSpot?
Yes. Sync leads, contacts, and conversation data directly into HubSpot.
Can I use InsertChat with Zendesk?
Yes. Ticket creation, handoffs, and syncing support conversations are all supported.
Do you support Shopify?
Yes. Your agent can answer product questions, check order status, and assist with common e-commerce queries.
What about WooCommerce?
Yes. WooCommerce works similarly to Shopify with access to product catalogs and order information.
Can I connect Google Workspace?
Yes. Connect Google Drive, Docs, and other Workspace tools as knowledge sources.
Do you have a Zapier integration?
Yes. Connect InsertChat with thousands of apps via Zapier to automate workflows and sync data.
Can the agent search the web?
Yes. Enable web search so the agent can find current information beyond your knowledge base.
Do you support calendar booking?
Yes. The agent can schedule meetings directly during conversations.
Can I use webhooks?
Yes. Send events to your own systems for custom integrations and real-time notifications.
Do you have an API?
Yes. Full REST API for creating agents, managing sources, and interacting with conversations programmatically.
Can I install it with Google Tag Manager?
Yes. Install via script embed or Google Tag Manager.
Can I embed it in my product?
Yes. Use in-app embeds for a native feel, or the API to build a custom interface.
Do you support custom SMTP?
Yes. Custom domain and SMTP options are available so outbound messaging aligns with your infrastructure.
How do I connect my first integration?
Start your trial, go to Settings > Integrations, and connect in one click. 600+ apps available.
Ready to build with GPT-OSS 20B?
Start your 7-day free trial today.
7-day free trial · Cancel anytime · No commitment