CATEGORY · 63 SERVICES

AI chatbots & assistants

Compare included quotas, restrictions and official provider documentation.

Free allowance

Free AI chat on web and mobile; writing, web search, files and code capabilities with variable usage limits

Keep in mind

Message and model limits vary with demand and reset windows; Claude Code and higher limits require paid plan. Commercial-use permissions for Free need confirmation.

Free allowance

Free AI chat, web search and limited access to file uploads, image tools and data analysis

Keep in mind

Message and tool limits vary; higher usage and advanced capabilities require paid plans. Consumer Free is not an API credit tier.

Free allowance

Free Gemini app with Google Account; AI chat, image creation, Gemini Live and limited Deep Research

Keep in mind

Model and feature access varies by region and changes with demand; paid Google AI plans offer higher usage. Not the Gemini Developer API.

Free allowance

Practically unlimited basic searches with citations; limited Pro Searches and basic file uploads

Keep in mind

Advanced models, image generation and premium features require paid plans; Pro Search quotas can change. Commercial-use eligibility must be checked.

Free allowance

Free web and mobile AI assistant with access to Mistral models, limited chat messages and web searches, limited coding sessions, basic image generation/editing, and 100+ connectors; usage allowances are not published as fixed numeric quotas.

Keep in mind

Free usage is rate-limited and subject to fair-use restrictions. Higher chat, search, image and coding limits require paid plans. Model training and commercial-use conditions should be reviewed before submitting sensitive or client data.

Free allowance

Free web and mobile AI chat for writing, brainstorming, summaries and real-time web answers; Edge page insights. Personal-account sign-in adds chat history, image creation and longer conversations. Free usage is capacity-limited, with no fixed public message quota.

Keep in mind

Usage limits, features and regional availability vary; Microsoft 365 business Copilot, in-app Office integrations and paid subscriptions are separate. Consumer Free commercial-use rights are not independently confirmed.

Free allowance

Free private AI chat with six available models and a recurring daily usage allowance; provider does not publish a fixed message or token quota.

Keep in mind

Daily free usage cap varies by model and anonymous abuse controls; higher limits and advanced models require a subscription. Model-provider usage policies apply; commercial-use eligibility is not explicitly confirmed.

Free allowance

Free AI chat with limited messages, Max-model usage and image generations; 1 Project; signed-in Free chat history retained for 7 days. No fixed numeric message quota published.

Keep in mind

Proton account required for saved Free-plan chats; guest chat works without signup but has tighter limits. Free usage caps are not numerically published. Lumo Plus and Lumo for Business are paid; commercial use of Free is not explicitly confirmed. Automated or bulk account creation is prohibited.

Free allowance

100 notebooks per user and 50 sources per notebook; free source-grounded AI chat and notebook outputs with compute-based usage limits refreshing every five hours.

Keep in mind

Formerly NotebookLM. Generation quotas vary by task complexity, chat length, source count and feature; higher limits and some advanced tools require paid Google AI/Workspace plans. Regional availability, output watermarks and commercial-use terms should be checked.

Free self-hosted software
Free allowance

MIT-licensed self-hosted multi-provider AI chat and agent interface, with Docker Compose deployment and no vendor-imposed seat, message, or monthly usage quota for the self-hosted software.

Keep in mind

Operator supplies server/container resources and model access; paid model API usage, third-party services and hosting may incur costs. LibreChat does not bundle free model API credits; self-hosted setup requires MongoDB and other services.

Free self-hosted software
Free allowance

Free MIT-licensed self-hosted Docker AI workspace with document chat/RAG, configurable LLM providers, agents and developer API; no vendor-imposed chat or document quota for the core software.

Keep in mind

Runs on your own compute and storage. Local model hardware, paid cloud LLM APIs, vector databases and managed AnythingLLM Cloud plans are separate costs. Secure internet-facing instances, review third-party model licenses and disable optional telemetry if desired.

Free allowance

Free Apache-2.0 desktop AI assistant for Windows, macOS and Linux; run supported open-weight language models locally and offline, or connect optional external model providers. No Jan software seat or message quota.

Keep in mind

Local inference requires sufficient CPU/GPU, RAM and disk space; downloaded model weights have separate licenses and commercial-use conditions. Optional hosted model APIs can charge fees. Attribution requested where appropriate.

Free allowance

Free MIT-licensed local AI chat desktop application and developer bindings; download supported models, chat privately offline and use LocalDocs or a local API server without GPT4All software message or token fees.

Keep in mind

Model weights have separate license and commercial-use terms; large downloads, RAM/CPU/GPU and disk space are required. Optional integrations and cloud services may have separate costs; local app is not a managed hosted API.

Free self-hosted software
Free allowance

Free MIT-licensed self-hosted document-grounded AI agents and assistants, knowledge ingestion, cited answers, workflows and local-model support; no software license or seat fee for self-hosting.

Keep in mind

Bring your own compute, storage, embeddings and language models. Paid cloud/managed/enterprise offerings and third-party model API usage are separate; review model and connector licenses and secure deployments.

Free allowance

Apache-2.0 open-source terminal AI pair-programming tool with repository mapping, Git integration and code editing; run with your own Ollama/local models without an Aider software subscription or vendor-imposed message quota.

Keep in mind

No free hosted model credits are bundled. Local models require suitable RAM/GPU and may produce lower-quality edits; paid LLM APIs, compute and third-party services cost extra. Review each model's separate commercial license and security before granting file/command access.

Free allowance

Apache-2.0 open-source AI coding agent for IDE, CLI and SDK, with code editing, terminal tools and support for locally hosted models; no subscription or software usage quota for the open-source agent.

Keep in mind

Requires a compatible model: hosted LLM APIs, optional Cline credits, local inference hardware and external tools may cost extra. Agent access to repositories and terminal commands requires careful approvals and sandboxing; enterprise/team services are separate paid products.

Free allowance

Free browser-integrated AI assistant for chat, summaries, writing and translation with rate-limited default and selectable models; no account required.

Keep in mind

Requires Brave browser; provider does not publish a fixed numerical free request allowance and models/rate limits can change. Premium subscription is separate; commercial-use terms require checking.

Free self-hosted software
Free allowance

AGPL-3.0 self-hosted AI second brain for searching personal documents, chatting with local or external LLMs, and creating custom agents; no software-license query or agent-count quota.

Keep in mind

Self-hosting requires your own compute, storage and optional model/API credentials; paid model APIs and hosted Khoj plans are separate. AGPL-3.0 network-use and source-sharing obligations apply.

Free self-hosted software
Free allowance

MIT-licensed OpenHands core AI coding agent, CLI, SDK and agent-server for local/self-managed use, without a software-license task quota.

Keep in mind

LLM provider usage and local compute can cost money. OpenHands Cloud/Enterprise and the enterprise/ directory use separate commercial or time-limited licenses; this listing covers only the MIT core.

Free allowance

Free AGPL-3.0 community desktop AI assistant supporting multiple model providers and local models; no vendor-imposed limit on locally running the community software.

Keep in mind

Cloud model API subscriptions, tokens and local compute may cost extra. Commercial use is allowed under AGPL-3.0 obligations; a separate commercial license is offered for exemption from AGPL terms.

Free self-hosted software
Free allowance

Free self-hosted browser interface for local and OpenAI-compatible AI models; no vendor-imposed usage quota when running the software with required branding intact.

Keep in mind

Current Open WebUI License (v0.6.6+) is source-available, not OSI-approved: keep Open WebUI branding intact unless the documented <=50-user exception or explicit permission/license applies. Hosting, model inference and third-party API costs are separate.

Free allowance

MIT-licensed self-hosted AI chat interface for OpenAI-compatible, local and cloud model endpoints; Docker/Node deployment with no NextChat software subscription or vendor-set chat quota.

Keep in mind

Self-hosted software is free, not the separately marketed hosted NextChat service. Cloud LLM API tokens, model inference, server/hosting and optional external services may cost money; secure access codes and API keys.

Free allowance

MIT-licensed multi-model AI chat workspace with personas, parallel Beam conversations and Docker/on-prem deployment; no vendor-imposed self-hosted software usage quota.

Keep in mind

Requires your own server or compatible deployment platform and API keys for paid LLM providers; model usage and infrastructure are not included. Hosted big-AGI subscriptions are separate; preserve MIT license notices.

Free allowance

Apache-2.0 open-source self-hosted AI agent and workflow builder with integrations, run logs and knowledge resources; run the software on your own infrastructure without a Sim Cloud subscription.

Keep in mind

Self-hosted production Docker Compose stack requires substantial RAM (project setup guidance notes 12 GB or more) and local secrets. Model API, hosting, storage and third-party integrations may incur fees; Sim Cloud is separate.

Free self-hosted software
Free allowance

Apache-2.0 open-source platform to build and deploy AI agents, workflows, plugins and knowledge bases on your own infrastructure using Docker Compose; no Coze Studio software license fee.

Keep in mind

Self-hosting requires at least 2 CPU cores, 4 GB RAM and Docker Compose. LLM API usage, hosting, databases and integrations can incur charges; Coze's managed platform is separate.

Free allowance

Free Community license for a self-hosted AI code-completion and assistant server, with Apache-2.0 core software, IDE extensions and support for locally run coding LLMs.

Keep in mind

Community features and seats differ from paid Team/Enterprise; repository ee/ code has a separate commercial license. You provide CPU/GPU, storage and model inference; model licenses and paid remote APIs are separate.

Free allowance

Open-source terminal AI agent; personal Google-account free access currently lists up to 1.000 model requests per user per day (capacity limits apply).

Keep in mind

Free access is for eligible individual Google accounts and may be restricted to Flash models; commercial eligibility for the free account must be checked. Paid API keys, Vertex AI billing, and higher tiers are separate.

Free allowance

MIT-licensed AI coding agent for VS Code, JetBrains and CLI; locally installed extension and CLI have no software seat or usage quota.

Keep in mind

AI inference is not included: hosted model tokens, Kilo gateway and cloud agent may be billed; use a compatible locally hosted model to avoid model API fees. Agent can execute commands; review permissions.

Free allowance

MIT-licensed local software-engineering agent that works on repository issues and coding tasks using a user-selected language model; no license-imposed usage quota.

Keep in mind

Software is free but hosted LLM/API tokens, local GPU/CPU, Docker execution and CI may incur costs; no bundled free model inference or managed agent service. Review sandbox permissions and model/provider terms.

Free allowance

MIT-licensed desktop visual IDE and TypeScript runtime for designing, debugging and running AI-agent/prompt-chain graphs locally; no Rivet software seat or workflow usage charge.

Keep in mind

LLM calls, embeddings, speech APIs, remote execution and storage are not included and may require paid third-party API keys. This is the Ironclad desktop IDE, not the separate Rivet Actors orchestration product.

Free allowance

MIT-licensed self-hosted AI answering engine with cited web results, bundled SearXNG search, and support for local Ollama models; no vendor-enforced software query quota.

Keep in mind

Run Docker and local LLMs on your own hardware; electricity, compute and hosting are not included. Optional OpenAI, Anthropic, Google or other hosted model APIs can incur separate charges. Formerly known as Perplexica; do not list both names separately.

Free allowance

MIT OR Apache-2.0 licensed Rust AI agent runtime for local use, with pluggable model providers including local Ollama; no software usage or seat limit.

Keep in mind

Runs on your own hardware; external LLM APIs, hosting and messaging services may cost extra. The agent can execute tools, so configure approval and sandbox policies; ZeroClaw name/logo trademarks are separate from the software license.

Free allowance

MIT-licensed self-hosted local-first AI assistant for notes, tasks, documents, photos and bookmarks; local-model inference and OpenAI-compatible API; no vendor-imposed software seat or request quota.

Keep in mind

Run and secure your own Docker/Compose stack, storage and model inference; requires sufficient CPU/GPU resources. Project warns it is not hardened for direct public exposure; use private networking/authentication. Optional external model APIs and infrastructure can incur charges.

Free allowance

Apache-2.0 open-source terminal, editor and desktop AI coding agent; free to install and run locally with user-selected local models or external model providers.

Keep in mind

Free software does not mean free inference: hosted Qwen/OpenAI/Anthropic/Gemini APIs and cloud resources may require payment or billing setup. Do not rely on discontinued Qwen OAuth promotional free quotas; local model hardware and model licenses are separate.

Free allowance

Apache-2.0 local desktop workspace for multiple coding agents, isolated task branches, approval gates, project management and scoped deployment credentials; free software on macOS, Windows and Linux.

Keep in mind

Windows/Linux support is beta; separate Claude, Codex, Gemini or other agent subscriptions, API calls, cloud resources and deployment accounts may incur charges. Free desktop software does not include paid agent/model access.

Free allowance

Apache-2.0 desktop AI-agent workspace for macOS, Windows and Linux; local Go sidecar, multi-model chat, knowledge base, tools and task automation; no software usage quota.

Keep in mind

Online LLM/API providers may charge separately; local Ollama models require device resources. Windows/Linux need separately installed Ollama for local models; pre-release builds and platform dependencies vary. Desktop software is free, not a free hosted inference tier.

Free allowance

Apache-2.0 open-source VS Code coding-agent extension with code, architect, ask and debug modes; no license fee for the locally installed extension.

Keep in mind

Zoo Code is the actively maintained community continuation of discontinued Roo Code, not Roo Code's former hosted service. Model/API tokens, cloud inference and external tools can incur separate charges; local models need your own hardware. No free hosted inference quota is included.

Free allowance

Apache-2.0 Neovim plugin for AI-assisted chat, inline editing and coding agents; no software license fee or plugin usage quota.

Keep in mind

Requires Neovim and an LLM provider, compatible local model or agent. Model API requests, subscriptions, agent tooling and compute are NOT included; local model hardware costs are yours. Not a vendor-hosted free inference plan.

Free allowance

Apache-2.0 self-hosted AI agent orchestration platform with Docker-based workspaces, agent schedules, approval gates, audit logs, and Claude Code/Codex/Gemini CLI runtime integrations; open-source core has no software license fee or vendor-hosted quota.

Keep in mind

Operator supplies Docker host, compute, storage, PostgreSQL and network infrastructure; model-provider API, agent runtime subscriptions, hosting and bandwidth may cost extra. Enterprise modules including organization SSO, advanced RBAC and compliance features require separate commercial licensing; do not treat them as included in the free open-source core.

Free allowance

MIT-licensed Rust AI agent runtime and local assistant with command-line interface, persistent memory, task planning, tools, MCP integration and optional messaging channels; self-run binary has no Ryvos software license fee or vendor-imposed local-use quota.

Keep in mind

Runs on user-controlled Linux/macOS hardware; LLM provider API calls, cloud servers, model subscriptions and optional Docker sandboxing infrastructure may cost extra. Ryvos Cloud is a separate hosted offering and is not included or verified as free; use local models where supported to avoid paid inference.

Free allowance

Apache-2.0 desktop cowork and AI-agent interface for macOS, Windows and Linux with local file workflows and support for connecting external or local AI agents; no license fee for the open-source app.

Keep in mind

Model subscriptions, API tokens, third-party coding agents and remote hosting may incur charges; the app does not grant free paid LLM inference. Grant filesystem and shell permissions carefully because connected agents can execute local operations.

Free allowance

Apache-2.0 self-hosted agentic workflow platform for AI orchestration, model/tool integration, RPA and team collaboration with Docker-based deployment; no open-source software license fee.

Keep in mind

Requires self-operated compute, containers, database, authentication and secure networking. Model APIs, external iFLYTEK cloud services and infrastructure can be billed; change default example credentials and review each integrated provider's terms.

Free allowance

MIT-licensed self-hosted encrypted inbox for AI agent artifacts, human approval of email drafts and passkey access; no software subscription, expiry or vendor-imposed self-hosting quota stated.

Keep in mind

Requires your own PostgreSQL, S3-compatible storage, HTTPS/WebAuthn origin and email verification delivery; infrastructure and outbound email can cost money. Self-hosted edition is distinct from arelay.app hosted plans. Upstream is young; review security and operational readiness. Default artifact/account limits may be enforced by application configuration.

Free allowance

GPL-3.0-licensed local autonomous AI assistant for browsing, coding and task planning; no software license fee or vendor-hosted quota

Keep in mind

Runs on your own computer with Python, Docker/Compose, local LLM and substantial GPU resources; electricity, hardware and optional paid model/API services are not included. Autonomous browser/code execution needs careful sandboxing; GPL-3.0 redistribution obligations apply.

Free self-hosted software
Free allowance

Apache-2.0-licensed agent framework and self-managed AgentOS runtime with agent/team workflows, API and MCP; no software license fee or vendor-hosted usage allowance

Keep in mind

You operate and pay for compute, PostgreSQL, storage and any model/provider APIs; optional Agno-hosted UI/integrations may have separate terms. Apache-2.0 attribution/license conditions apply; no free cloud hosting or model tokens are included.

Free allowance

Free MIT-licensed local-first Rust AI coding agent with desktop and CLI modes; no software subscription, usage quota or expiration published. Supports local models and bring-your-own model providers.

Keep in mind

Windows x64 v0.1.0-rc.1 is an unsigned pre-release; production maturity is unproven. Model API calls, local hardware and remote inference are not included; paid provider billing may apply. Review agent command/file permissions and retain backups. License requires attribution to Argentum by AG064.

Free allowance

MIT-licensed local-first MCP coordination server with durable SQLite agent inboxes, tasks, channels, and optional HTTP/SSE integration; no software subscription or hosted quota.

Keep in mind

Free software only, not hosted AI inference or compute. Bring your own agent clients, model subscriptions/API usage, and infrastructure. Default local-trust HTTP endpoints should not be exposed on shared/public networks; configure dashboard and HTTP secrets for remote/team use. Some federation features are roadmap-only.

Free allowance

MIT-licensed self-hosted agent orchestration dashboard with task dispatch, agent monitoring, scheduling, RBAC, audit logs and SQLite storage; no software usage quota published.

Keep in mind

Alpha software; self-hosting and agent/model provider usage are not included. Review security hardening, authentication and external gateway/API costs before production use.

Free self-hosted software
Free allowance

Apache-2.0 open-source AI customer support chatbot with shared inbox, human handoff, web chat widget, knowledge-base answers and bring-your-own LLM; self-hosted core has no provider-imposed usage quota.

Keep in mind

Requires your own server, database and model/API provider; LLM tokens and messaging-channel services can cost money. Enterprise features and support are separately licensed. Production security and external integrations require configuration; no free cloud infrastructure is included.

Free allowance

Free desktop and self-hosted AI bot workspace for personal noncommercial use without a stated time limit or bot-count cap. Includes multiple bots with individual memory and computer/browser access, scheduled and event-triggered routines, connected apps and MCP integrations on the user's own machine or server.

Keep in mind

Licensed under PolyForm Noncommercial 1.0.0 (source-available, not an OSI-approved open-source license); commercial use requires a separate license. Free software does not include AI inference, an AI subscription, paid API usage, Dusk Cloud or server costs. Local routines stop when the computer sleeps; running continuously requires an always-on self-managed server or paid hosting.

Free allowance

Ad-supported Free coding agent for desktop, CLI and web, with unlimited messages per hour-long session and daily hours that reset at local midnight. Official November 7 guide says 5 hours/day in countries outside listed full-access regions (4 via VPN); country, model multipliers and account eligibility affect effective hours.

Keep in mind

Homepage says unlimited, but free usage is daily-hour-limited and model-dependent; /plans advertises full-access figures that do not apply globally. Free access requires a human present, one account per person, and may require phone or identity verification. No direct API, scripted/bot use, resale or account sharing; terms and quotas may change. Ad-supported; code and prompts are processed by the service/model providers. Cloud inference is included, not a self-hosted free inference service.

Free allowance

Free personal AI agent for web, iOS and Android with a usage allowance that refreshes after the free limit is reached; Meta does not publish a fixed free token quota on its product overview.

Keep in mind

Availability is region-gated; Meta has announced launches in the US, Japan and Singapore, while other accounts may encounter a waitlist. Higher usage requires a paid subscription. One-time referral token bonuses are separate from the ongoing free allowance and are not included in this listing.

Free allowance

MIT-licensed self-hosted AI workspace software for chat, agents, documents, research and local model management; no software usage quota or expiration stated.

Keep in mind

Bring your own hardware and models; optional external model APIs and integrations may incur charges. Secure authentication and network access before deployment; third-party components retain their own licenses.

Monthly credit
Free allowance

Free Starter includes 20.000 PT (up to US$20 of Perch-hosted model usage) per calendar month, resetting in UTC; Web, Desktop and CLI access.

Keep in mind

No direct public API access; hosted models and limits may change. Starter Data Program is enabled by default but can be disabled in Privacy settings. No overage charges; usage may pause or be throttled at limits. BYOK inference costs are separate. Commercial-use permission for Starter is not explicitly confirmed.

Free self-hosted software
Free allowance

Apache-2.0 self-hosted AI agent control plane with MCP firewall, model gateway, policy rules, human approvals, agent flows, private runners and runtime observability; no published software usage quota or expiry.

Keep in mind

Self-hosted OSS does not include free compute, model API calls or managed cloud runners. Team RBAC, account-level role controls and some enterprise features require Cloud/Enterprise; self-hosted operators must secure credentials and infrastructure.

Free local software
Free allowance

Free, open-source MIT-licensed local-first desktop workspace for up to 10 supported coding-agent runtimes, with Git worktrees, diffs, browser verification, task handoffs and pull-request workflows on your own computer. No Synara account, payment card or fixed free-plan expiration is required for local use.

Keep in mind

Synara Free is local software, not a hosted AI model or free inference service. Install and authenticate supported coding-agent CLIs separately; provider subscriptions, API charges, quotas and permissions are not included. Synara Alto remote control between computers, approvals/alerts and the online profile require a paid subscription; iPhone/iPad remote support is marked coming soon.

Free local software
Free allowance

Free local AI-agent orchestration workstation for one user on one machine, including agent missions, Kanban, git worktrees, code/review workflows, routines, local memory, built-in browser and plugins; works with separately provided agent CLI subscriptions, credentials or endpoints.

Keep in mind

Timon Free is local software, not free hosted AI inference or compute. Claude Code, Codex, Copilot and other model CLI subscriptions/API usage must be separately provided and may cost money. Free is limited to one user and one machine, with no support; multi-user roles, premium plugins and server-based Cloud edition require a paid plan. Main Timon code is AGPL-3.0-only with a plugin exception; distributing modified software or providing a modified network service may trigger source-sharing obligations. Pro/Cloud pricing has not been announced.

Free allowance

Apache-2.0-licensed self-hosted AI agent collaboration platform with control plane, worker daemons, task/workspace orchestration, isolated Git worktrees, BYOK model configuration and self-managed cloud-node runtimes; no software usage quota or expiry stated.

Keep in mind

Free software does not include hosted model gateway, managed cloud-node pool, subscription credits, VPS, PostgreSQL/object storage, or AI model/API usage. Operate and secure your own workers and credentials; commercial hosted services are separate.

Free allowance

MIT-licensed self-hosted browser interface for existing Claude Code and Codex CLI sessions, including mobile approvals, session management, file/diff review and notifications; no software seat quota or expiry stated.

Keep in mind

Requires user-provided machine, Node.js and installed/authenticated Claude Code or Codex CLI. Agent subscriptions or API usage are not included and may cost money. Optional public relay is third-party infrastructure; direct self-hosted access requires secure networking and authentication. Desktop installers are beta.

Compare selected services
← Browse services
FREE TIER COMPARISON

Compare free services

Choose 2–4 services and see their allowances, billing requirements and restrictions side by side.

Select at least two services to compare.

Use the search above or select Compare on service cards in the directory.

Explore the directory
Donate
Support Free Tier Hub

Help keep this independent free-tier directory available and up to date.

Donate via ePlus.DEV ↗GitHub Sponsors ↗Completely optional. Donations never affect which services we list or recommend.