Introduction
Welcome
Abundly is an enterprise-grade platform to build, configure, manage, and collaborate with autonomous AI agents — an operating system where teams of humans and AI agents work side by side. Unlike simple tool-enabled AI chat, Abundly agents offer true autonomy (schedules, triggers, 24/7 operation), multi-modal communication (SMS, email, Slack, voice calls), conversational configuration, multi-agent collaboration, dynamic apps and scripts, and enterprise security and governance. Example use cases include cross-platform automation (“dig through last week’s Slack messages and create a Notion overview doc”), scheduled monitoring (“every morning, check my HubSpot todos and ping me on Slack if something is urgent”), and bulk data processing across spreadsheets and integrations. Full page →What is an AI agent?
An Abundly agent is a multilingual, autonomous digital colleague — faster and cheaper than humans, more intelligent and flexible than code, sitting in between. Agents are designed to augment humans by handling routine knowledge work that involves fuzzy inputs, uncertainty, and judgment, freeing people for strategic tasks. Every agent has four components: an LLM (Claude, GPT, Gemini, or others) as its brain, a Mission (natural-language instructions that are versioned and can be self-updated), Tools (grouped into capabilities like Slack, Gmail, code execution, web search, and 40+ integrations), and Autonomy (schedules, triggers, and proactive communication). Benefits over manual work include lower cost, higher speed, better control through written instructions, and higher consistency. Full page →Getting started
Sign up at app.abundly.ai for a free trial with credits — no credit card required. Create an agent from scratch or clone one from the Agent Catalogue. Configure your agent by chatting with it (it writes its own instructions from the conversation), enable the capabilities it needs, then test and refine. Full page →Platform overview
The Team page is your home base, showing all agents as cards with access to Members, Agent Management, and Usage & Limits in the sidebar. The Agent page is where you configure and interact with an individual agent — it includes Instructions, Docs, Capabilities, Diary & Approvals, and Settings (covering communication, usage limits, agent-to-agent communication, team access, API endpoint, model selection, and more). The chat interface supports text, voice input, file uploads, Walk & Talk mode, fact checking, and a thinking toggle for more deliberate reasoning. Full page →Pricing
Abundly uses a credit-based pricing model with no per-seat charges — you get unlimited agents and unlimited team members, and pay only for credits consumed when agents perform actions. Three subscription tiers are available: Start (1 user), Pro (up to 4 users), and Enterprise (custom). Each subscription includes monthly credits, and unused credits roll over for up to 12 months. You can set per-agent spending limits and track usage through the Usage Reports dashboard. Full page →Features
Agent instructions
Instructions are the foundation of your agent’s behavior—a working agreement describing what the agent should do, how it should behave, and the context it needs. A new agent starts with empty instructions, making it a general-purpose chatbot; adding instructions gives it a purpose. You can edit instructions directly or configure them conversationally by chatting with the agent and asking it to draft its own instructions based on the discussion. Instructions are versioned automatically, letting you track changes, compare versions, and revert if needed. Full page →Configurable capabilities
A capability is a feature or integration your agent can use—such as generating images, executing code, or connecting to Slack and Google Drive. Agents start with a basic set and you add more as needed. If you ask the agent to do something requiring a missing capability, it automatically requests it. Capabilities can be combined in powerful ways, for example pairing Web Search + Edit Documents + Send Email to automate research workflows. Full page →Communication, chat, email, SMS, and voice
Agents communicate through multiple channels: built-in chat, email, SMS (beta), voice calls (beta), and collaboration tools like Slack and Teams. By default only chat is enabled—you decide which channels to allow. The chat interface supports voice input, file uploads, interactive apps, code execution, rich responses (formatted text, diagrams, images, voiceovers), and multi-user collaboration. Each conversation acts as a local context; the agent also has global context from instructions and documents. Every agent gets its own email address (e.g.,myagent.myteam@agent.abundly.ai). Email supports attachments, CC/BCC, rich HTML formatting, and configurable guardrails including recipient whitelists and approval requirements. SMS works similarly but is currently in beta.
Voice features include microphone-based voice input transcribed to text, a text-to-speech speak button, and Walk & Talk mode for mostly hands-free interaction. Phone calls (beta) let agents make and receive calls using real-time voice models. For complex requests, the “receptionist pattern” captures the problem during the call and processes it afterward with the agent’s full intelligence. Voice generation is powered by ElevenLabs and OpenAI.
Communication → · Chat → · Email & SMS → · Voice →
Documents, databases, apps, and scripts
Each agent has its own file repository with version history. Supported file types include text, Markdown, code files, SVG, PDFs, Word, PowerPoint, spreadsheets, images, audio, and video—non-text files are automatically transcribed or described. Documents have configurable visibility levels: Full (always in context), Summary (default—agent reads on demand), Searchable (found via search only), and Hidden. Documents can be published with a public URL. Agent databases are lightweight structured data stores (JSON records, like a mini MongoDB) that agents can create instantly through conversation. They support direct queries, semantic search via RAG, and schema management. Databases pair with interactive apps for visual browsing and editing. For larger-scale needs, external databases can be connected via MCP. Interactive apps are agent documents containing executable React or HTML code. Agents build dashboards, forms, calculators, and custom tools on demand. Apps can be published publicly, and when backed by a database, you configure read/create/update/delete permissions for public viewers. Code execution lets agents write and run JavaScript for faster, cheaper, and more reliable data processing compared to token-by-token AI reasoning. Scripts can access all agent capabilities (web search, databases, email) and can be saved as reusable agent documents and scheduled to run automatically. Documents → · Databases → · Apps → · Scripts →Scheduled tasks
Agents can wake themselves at specific times to act proactively. Recurring tasks are defined in the agent’s instructions and run on any pattern (daily, weekly, monthly). One-off tasks are scheduled through chat and run once. You can test scheduled tasks immediately by asking the agent to simulate the trigger. Tasks can be enabled, disabled, or managed from Settings → Scheduled Tasks. Full page →Context and memory
Global context (instructions and documents) is always available to the agent. Local context is per-interaction: chat conversations are independent, and triggers include only the relevant event data. The Memory capability provides a dedicated semantic search store for facts and preferences the agent accumulates over time—finding memories by meaning, not exact text. You can also use instructions, documents, or agent databases as persistent memory. The agent can read its own diary for historical context. Full page →Web search, scraping, and deep research
Three built-in search capabilities require no API keys. Web Search uses Perplexity for quick lookups returning results in seconds with citations. Web Scraping reads live page content from specific URLs. Deep Research conducts comprehensive multi-source analysis taking 10+ minutes but delivering results comparable to days of manual work. Web search reduces hallucination and provides verifiable sources with citation URLs. Full page →Multi-agent collaboration and task delegation
Agents can collaborate via delegation (creating temporary sub-agents) or agent-to-agent communication (persistent agents exchanging messages). Delegation is primarily for context management—sub-agents handle focused tasks with clean contexts and return concise results, keeping the parent agent efficient. Agent-to-agent communication connects specialized agents you’ve already configured, enabling reusable expertise across multiple contexts. An alternative to multi-agent setups is task documents—a single agent with separate documents describing different workflows. Full page →Model selection
You choose which LLM powers your agent—Claude, GPT, or Gemini—or leave it on “(no preference)” for Abundly’s recommended default. Model aliases (e.g., “Claude Sonnet Latest”) auto-update to verified new versions; specific versions lock behavior. Available models range from high-capability options like Claude Opus 4.5, Gemini Pro 3, and GPT 5.2 Pro to cost-effective options like Claude Haiku 4.5, Gemini Flash Lite 2.5, and GPT 5 Nano. You can switch models mid-conversation and enable thinking tokens for Claude models to improve complex reasoning. Full page →Agent personalization
You can customize your agent’s profile picture (auto-generated, custom prompt, or uploaded), communication style via personality instructions, chat start message, description, and tags. Personality instructions shape tone and style—from formal to casual to humorous—and carry through to diary entries. Agents can also update their own name, profile picture, and description on request. Full page →Agent evals
Evals let you define test cases that verify your agent behaves correctly after changes to instructions, models, or capabilities. Each eval has a trigger prompt, optional tool overrides (include, exclude, or fake tools), and a grader that evaluates the response as pass/fail or a 1–10 rating. Validation rules provide fast deterministic checks on tool usage before grading. You can run evals across multiple models to compare performance side-by-side, export/import eval suites as JSON, and even let the agent create and run its own evals. Full page →Activity monitoring
The agent diary is a high-level record of activities and reasoning, written automatically when processing triggers and during significant chat events. The activity log provides a detailed technical view of each trigger’s processing pipeline: prefilter, interpretation, plan, security assessment, and execution with specific tool calls. Both tools help you understand why the agent made specific decisions, enabling targeted instruction improvements. Full page →Usage and limits
The Usage & Limits page gives administrators real-time visibility into credit consumption across agents, with balance history graphs and top-consumer breakdowns. You can set a default daily credit limit for all agents and override it per agent. When an agent hits its limit, it pauses until UTC midnight. Both LLM tokens and tool usage count toward credits. Full page →Team members and groups
Organizations support three roles: Admin (full control), Member (standard access), and Guest (invitation-only access to specific agents). Admins invite members, assign roles, and manage access. Groups (enterprise feature) partition agents and users into logical units by team or department, with delegated administration and per-group capability settings. Each group has its own admin, member, and guest roles. Capabilities disabled at the organization level cannot be enabled at the group level. Team members → · Groups →Agent management
The Agent Management page gives administrators a centralized view of all agents, their status, capabilities, privacy settings, and usage. Admins can pause agents, manage per-agent administrators, and control organization-wide capability defaults (default on or default off for new capabilities). When groups are enabled, group-level overrides let different teams have different capability configurations. Full page →Agent sharing and cloning
Internal cloning duplicates an agent within your workspace or to another team, copying instructions, capabilities, model settings, and optionally documents. Public sharing generates a link letting anyone clone your agent to their own workspace—visitors can see the agent’s name, description, and instructions. Cloned agents are fully independent from the original. Full page →HTTP requests and agent API endpoint
The HTTP capability lets your agent call any REST API on the internet—useful when no dedicated capability or MCP server exists for a service. The agent can discover APIs via web search, read documentation, and learn through trial and error. Approval can be required for every request. This capability is hidden by default and enabled on request. The Agent API Endpoint exposes your agent as a service by creating API keys that allow external systems to send HTTP requests directly to it—enabling integration with external agents and applications. HTTP requests → · Agent API endpoint →Integrations
Integrations overview
You give agents access to external services like Slack, Google Drive, and GitHub by toggling capabilities on and off. Some integrations use platform-provided credentials and work out of the box, while others require you to authorize your own account via OAuth or API token. Agents can request capabilities on demand during a conversation if they need something that hasn’t been configured yet. User approval can be required before an agent executes certain actions (e.g., reviewing outgoing emails). You can also choose which LLM powers your agent — Claude (Anthropic), Gemini (Google), or GPT (OpenAI) — under Settings → Advanced Config. If a built-in integration doesn’t exist for your use case, you can extend your agent via MCP servers or the HTTP capability. Full page →MCP servers
MCP (Model Context Protocol) is an open protocol for connecting AI applications to external tools. You can give your agent access to any MCP server by pasting the server URL into the MCP Servers tab — no platform update required. This lets you extend beyond built-in integrations, use community servers, or connect internal systems like a CRM or custom database. The platform auto-detects authentication requirements and supports OAuth or API token. Once connected, you can toggle individual tools on or off. MCP credentials are encrypted using hybrid AES+RSA encryption, and OAuth tokens are never exposed to the AI model directly. Only add servers from sources you trust, since an MCP server can perform actions on your behalf. Full page →Integration pricing
Billing depends on whose credentials are used. If you bring your own account (e.g., Slack, Notion), Abundly doesn’t charge — you’re paying the provider directly. If Abundly provides the account (e.g., web search, image generation), the cost is converted to credits and charged from your balance. LLM inference also consumes credits regardless of integrations. Full page →Anthropic
Anthropic provides the Claude family of language models, known for strong reasoning and safety features. Models range from fast responses to complex reasoning tasks and are configurable under Settings → Advanced Config. Full page →ElevenLabs
ElevenLabs provides high-quality text-to-speech for your agent, available alongside OpenAI TTS under the Text to Speech capability. Use it to generate voiceovers, audio summaries, or narrated reports — for example, “Generate a voiceover for this script using a male Scottish voice.” Full page →Firecrawl
Firecrawl enables your agent to read and extract content from web pages. It is enabled by default under the Web Scraping capability. Combine it with Perplexity to find relevant URLs first, then extract detailed content. Full page →Giphy
Giphy lets your agent search for and share animated GIFs, included alongside AI image generation under the Image Generation capability. Full page →GitHub
Connect your agent to GitHub to browse files, create and review pull requests, manage issues, and react to repository events via webhooks. Requires a personal access token. Example: “When a new issue is created, analyze it and add appropriate labels.” Full page →Google Calendar
Connect your agent to Google Calendar to view events, create meetings, and manage schedules. Currently hidden — contact support@abundly.ai to enable. Example: “Every morning, email me a summary of today’s client meetings.” Full page →Google Drive
Your agent gets secure access to specific Drive files you explicitly share — it can read, edit, and create documents and spreadsheets. The agent uses a narrow permission scope and cannot browse your entire Drive. Example: “Read the data from this spreadsheet and give me the key insights.” Full page →Google Gemini
Google provides Gemini language models as an alternative to the default, plus image generation via Gemini Pro 3 (NanoBanana). Enable Image Generation and select Google to generate images from text descriptions. Full page →HubSpot
Connect your agent to HubSpot to search contacts, companies, and deals, and create tasks or log emails. Currently hidden — contact support@abundly.ai to enable. Requires a HubSpot private app access token. Example: “Every morning, check for deals stuck in the same stage for over a week.” Full page →Mailgun
Mailgun is the default, platform-provided email delivery provider for the Send Email capability — no setup required. Emails are sent from your agent’s dedicated email address. You can configure recipient whitelists and approval requirements. Full page →Notion
Connect your agent to Notion to read pages, query databases, and create or update content. Requires an internal integration token, and the agent can only access pages you explicitly share with the integration. Example: “When I forward you an email from a client, add a new entry to our CRM database in Notion.” Full page →OpenAI
OpenAI provides GPT language models, image generation (GPT image), voice transcription (Whisper), text-to-speech, the Realtime voice API for phone calls (used with Twilio), and image recognition. Voice transcription works via the mic button or by dropping audio files; image recognition works on images provided through chat, documents, or email attachments. Full page →Outlook
Connect your agent to Outlook to read and search Microsoft 365 emails and create drafts. Currently hidden — contact support@abundly.ai to enable. Note: Outlook can read emails and create drafts but does not send emails directly. Full page →Perplexity
Perplexity powers your agent’s Web Search and Deep Research capabilities — platform-provided, no setup required. Web Search returns answers in seconds for quick facts and lookups; Deep Research takes longer but provides comprehensive analysis with extensive citations. Full page →SendPro
SendPro is an alternative email delivery provider for the Send Email capability, useful when you want emails to come from your own domain instead of the platform default (Mailgun). Requires your own SendPro Client ID, Client Secret, Account ID, and sending domain. Full page →SharePoint
Connect your agent to SharePoint to list, read, create, and edit documents in Microsoft 365 document libraries. Currently hidden — contact support@abundly.ai to enable. Works with text-based files; Office documents can be read but not edited directly. Full page →Slack
Your agent gets its own Slack identity and can search, read, respond to, and post messages in your workspace. Each agent needs its own Slack app with a bot token and app ID. By default the agent responds to @mentions; you can configure it to respond to all channel messages. Example: “Whenever a new potential customer is mentioned on #sales, research how they fit our ideal customer profile.” Full page →Trello
Connect your agent to Trello to manage cards, respond to board events, and automate workflows. Supports webhook events for card creation, movement, and comments. Example: “When a new card is created in the Inbox list, analyze it and move it to the appropriate list based on priority.” Full page →Twilio
Twilio powers SMS and phone call capabilities — Send SMS, Receive SMS, Make Phone Call, and Receive Phone Call. For phone calls, the platform uses Twilio together with the OpenAI Realtime API for natural voice conversations. Currently hidden — contact support@abundly.ai to enable. You can configure phone number whitelists and require approval before sending messages or making calls. Full page →Twitter / X
Connect your agent to Twitter/X for searching tweets, analyzing users, and posting content. Twitter/X Search uses the TwitterAPI.io service; Twitter/X Post requires OAuth credentials from an X Developer account with read-and-write permissions. Example: “Every morning, search Twitter for mentions of our company and post a sentiment summary to Slack.” Full page →Guides
Find your use case
Not every task benefits from an agent — some are better handled by code, others by humans. The sweet spot for agents lies in tasks that are routine and time-consuming, don’t require deep expertise or creativity, involve fuzzy inputs that pure code can’t handle, and carry correctable (not mission-critical) stakes. To find candidates, map how your team spends time along two axes: frequency/time investment and value of your time. Tasks that are frequent but feel like poor use of a human’s time are your best starting points. Good examples include screening incoming emails, generating weekly summaries from multiple sources, and reviewing documents against a checklist. Start with simple use cases to build intuition for how agents think and what context they need before tackling complex workflows. Full page →Agent design
Agent design is the process of figuring out what an agent should do, how it does its work, how success is measured, and how it interacts with humans and systems. The Agent Design Canvas is a framework covering ten aspects: Purpose, Triggers, Action Plan, Interfaces, Capabilities, Impact, Input, Knowledge & State, Output, and Success. The Action Plan is the heart of the design — numbered steps showing the workflow with clear human-AI handoffs. Apply the principle of least privilege when choosing capabilities, and define success in measurable business terms. Every agent needs a clearly defined human owner responsible for its behavior, analogous to an editor-in-chief who sets guardrails and follows up when things go wrong. Prototype early with fake data or simulated integrations — a useful first prototype can often be built in 30 minutes. Keep instructions clean by limiting context to what the agent actually needs, and plan for iteration: getting an agent from “works OK” to “awesome” typically takes a few rounds of refinement. Full page →Agent optimization
Every agent consumes credits when it works, with costs varying by task complexity and LLM choice. A simple news check might use around 50 credits, while screening 300 acquisition targets could use 5,000. To reduce costs, evaluate run frequency, trim context to essentials, store reference information in agent documents (with Summary visibility) rather than instructions, simplify instruction logic, and consider switching to a cheaper model if your use case allows it. Use the Usage & Limits dashboard to track per-agent consumption and set daily limits to prevent runaway costs. Full page →Troubleshooting
Common issues include “Overloaded” errors from LLM provider traffic spikes (wait and retry, or temporarily switch models), integration failures from missing permissions or expired tokens, and invalid verification codes caused by requesting multiple codes (only the most recent code is valid, and codes expire after 30 minutes). Full page →Use Cases
Use Cases
Abundly publishes detailed use cases at abundly.ai/use-cases, each showing the challenge, solution, and results. Below is a categorized summary. Support & knowledge base:- HR Support Agent — Answers employee questions about policies and processes, escalates sensitive issues to HR colleagues, and keeps the knowledge base current.
- Helpdesk Agent — Handles incoming support questions with consistent answers, escalates complex issues with context, and learns from every interaction.
- Lead Generation Agent — Searches the web for companies matching your criteria, verifies relevance, and delivers contact details.
- Lead Analysis Agent — Researches incoming leads, scores them against your ideal customer profile, and alerts your team to high-potential opportunities.
- Sales Proposal Agent — Transcribes meeting notes, drafts tailored proposals, and formats everything to your company template.
- Commercial Contracts Agent — Organizes your contracts, answers questions about terms and pricing, and enables analysis across your entire contract portfolio.
- Procurement Agent — Reviews procurement documentation against standards, validates compliance, and flags contradictions.
- Target Screening Agent — Researches potential investment targets, scores them against your criteria, and surfaces the most promising opportunities.
- Security Vulnerability Agent — Monitors vulnerability disclosures, filters for relevance to your codebase, and alerts the right people.
- Product Updates Agent — Monitors your repository, translates commits into plain-language release notes, and posts them to Slack.