AutoGPT vs SuperAGI (2026)

Block categories span an MCP tools block, a Claude Code block, an E2B sandbox block, image, video and music generation via Ideogram, Flux, VEO3, Fal, Replicate and MusicGen, a generic HTTP Request block with multipart upload, host-scoped authentication and safe-redirect protection, a read-only SQL Query block, Reddit blocks via PRAW, and Pinecone and Mem0 blocks. Automations import from n8n, Make.com and Zapier.

A Block Development SDK lets customers build, register and ship custom blocks using a provider pattern, with auto-generated block documentation surfaced in builder search. A Run Agent block executes another agent as a sub-agent, and library operation blocks add agents from the store programmatically. Through the MCP tools block, the platform calls out to other systems.

Sourceagpt.co pricing page, docs.agpt.co and agpt.co block documentationread 2026-08-31

Toolkits let agents take authenticated action in outside systems. Search runs through Google Search, DuckDuckGo, Searx and SerpAPI, and developer toolkits cover GitHub, including Fetch PR and Review PR, Jira and a Coding toolkit.

Productivity toolkits cover Notion, Google Calendar and a File Manager, Email sends messages with attachments, Twitter and Instagram cover social posting, DALL-E and Stable Diffusion generate images, Apollo.io supplies sales data and Google Analytics supplies analytics. Credentials are set per toolkit in the GUI, with OAuth where the service requires it.

Developers can add custom toolkits through a linked GitHub repository, a cookie-cutter template and a community tools repository, so the catalog is open to additions.

Sourcegithub.com/TransformerOptimus/SuperAGI README and release notes v0.0.9 through v0.0.12read 2026-09-01

A Run Agent block runs another agent as a sub-agent within a workflow, taking the agent's graph ID, version and input data and returning its outputs, with input and output schemas defining the data passed between parent and child, for modular workflows, specialized agents called from a main orchestration agent, and dynamic routing by input type.

The visual builder composes blocks with branching, looping and routing logic, and AI reasoning blocks include an AI Condition block that lets a model decide a branch. Agents run continuously and on schedules, with execution, scheduling and notification managers in the backend. How failed sub-agents are retried, and whether branches run in parallel, is not described.

Sourceagpt.co and docs.agpt.co block documentation, agpt.co home and pricing pagesread 2026-08-31

Agents run a ReAct reasoning-and-acting loop, and the runtime handles the control flow rather than leaving it to the developer. Predefined agent workflows can be chosen at provisioning as sequential instruction sets, and Sales Engagement and Recruitment ship as worked multi-step examples. A wait block lets a workflow pause for a set time before continuing.

A task queue sets priorities, a configurable max-iteration limit caps each run, looping detection heuristics flag an agent that is stuck, and provider errors such as rate limits, authentication failures and exceeded quotas are shown to the operator instead of failing silently. Several agents can run at the same time.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.10 and v0.0.13, and README architecture diagramsread 2026-09-01

Webhook blocks give an inbound event route, scheduling blocks give recurrence with a dedicated scheduling manager in the backend, and a public REST API exposes agent execution programmatically. Agents run continuously and trigger from external sources, with execution, scheduling and notification managers and websockets for real-time communication in the backend. A sub-agent block allows one agent to start another. No chat, email, Slack or messaging channel through which a person addresses an agent is offered beyond the platform's own interface.

Sourceagpt.co misc block documentation, agpt.co home and pricing pages, docs.agpt.co platform documentationread 2026-08-31

An agent can be scheduled for a single future run or for a recurring run at a set interval, with optional expiry by deadline or by total run count, and schedules can be stopped, edited and saved, so the agent starts without anyone typing.

Webhooks fire outward when an agent is running, paused, completed, terminated or reaches max iterations, which notifies external systems but does not start work. People reach the agent only through the framework's own web interface at localhost or the hosted console, and there is no inbound channel from email, chat or a team tool. Email, Twitter and Instagram are outbound toolkits the agent acts through, not channels that reach it.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.8 and v0.0.12read 2026-09-01

A Vector and Memory block category covers Pinecone vector storage and Mem0 long-term memory, connecting agents to a vector index the customer keeps and queries.

Data transformation blocks include a spreadsheet reader, XML parser, text encoding and replacement and list and dictionary manipulation, an HTTP Request block supports multipart upload for document retrieval, and a read-only SQL Query block reaches the customer's own platform data. Mem0 blocks separately support knowledge building by accumulating facts over time. AI reasoning blocks include a Fact Checker.

The customer supplies the Pinecone index, since AutoGPT bundles no knowledge base of its own, and no chunking strategy, embedding model, retrieval configuration or citation behavior is described.

Sourceagpt.co pricing page, docs.agpt.co platform blocks documentationread 2026-08-31

Agents can work from the customer's own content, retrieved at run time. Files in pdf, docx, pptx, csv, txt and epub are uploaded through the resource manager, turned into vector embeddings by the LlamaIndex integration and queried by the agent through the Resource Query tool.

Knowledge Embeddings can also be created from the customer's own sources or installed from the marketplace, hosted in a Pinecone, Qdrant, Weaviate or Redis index the customer connects by URL, and searched semantically at run time through the KnowledgeSearch toolkit. A Knowledge Console logs embedding performance, queries and token consumption.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.7, v0.0.9 and v0.0.11, superagi.com knowledge-embeddings articleread 2026-09-01

Mem0 memory blocks let an agent add memories that persist across workflow executions, optionally scoped to the run or the agent and tagged with metadata, retrieve the latest memory or all memories with the same scoping, and search memory, with context review, memory export and listing memories to find outdated or incorrect information among the uses. Memory controls let the user review and forget stored facts and preferences when memory is enabled, and a persistent key-value storage block is shared across tasks and workflows. How long memories are kept is not stated.

Sourceagpt.co basic blocks documentation and pricing pageread 2026-09-29

Short-term memory is a rolling window sized to the model's token limit. Context that falls outside it is condensed into a Long Term Summary, and the two are combined into an Agent Summary that goes into each reasoning step, weighted seventy-five to twenty-five. Long-term memory persists in a vector database, with Qdrant available as an LTM backend alongside Pinecone, Weaviate and Redis. Tool Memory stores every tool response from a run as vectorized chunks in Redis for semantic retrieval, though only the Thinking Tool can read it.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.8, v0.0.10 and v0.0.11read 2026-09-01

A Human In The Loop block pauses workflow execution and presents data to a human reviewer for approval, waiting until the reviewer approves or rejects it and routing the data to separate approved and rejected outputs, with an option letting the reviewer edit the data before deciding and an optional review message explaining the decision.

Uses include reviewing AI-generated content before publishing and requiring manager approval for purchases, access requests or configuration changes. Approval gates can be placed anywhere in a workflow, including nested workflows, with execution pausing for review and then resuming. Guardrails beside the approval step include per-task LLM budget caps, a spending cap per Expert, host-scoped HTTP credentials and safe-redirect protection.

Sourceagpt.co basic blocks documentation and pricing pageread 2026-09-29

Restricted mode is chosen per agent as a permission type at provisioning, and an agent running under it must get human approval before critical actions such as writing files, sending emails and moving to the next iteration. Requests appear in the Action Console, where the operator can Approve, Give feedback or Deny, so the customer decides which kinds of action need sign-off, who gives it and what happens while the request waits. A configurable max-iteration limit caps a run separately from approvals.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.4, v0.0.6 and v0.0.10read 2026-09-01

Access controls include OAuth 2.0 based single sign on, Google sign-in, API key generation, rotation and revocation, a credentials manager with scope filtering, host-scoped HTTP credentials, OAuth callback hardening and token validation, DNS-rebinding and safe-redirect protection, ClamAV scanning of uploaded files, and admin impersonation with audit trails.

Admin controls and roles and security and compliance options are listed for a Team plan marked coming soon. No security certification, attestation or trust page is published, and no SAML, SCIM or role based access control is available on the current plans.

Sourceagpt.co pricing page and docs indexread 2026-09-29

Authentication runs through GitHub OAuth, which also protects the APIs, alongside a configurable auth secret and an option to restrict list creation, and the project runs CodeQL code scanning in its own pipeline.

The framework has no access model for who inside an organization can make an agent do what, and it publishes no compliance page or certification of its own. Its logs record telemetry, and SuperAGI names no audit trail. The GDPR, SOC 2 and HIPAA claims made for the company's separate commercial platform belong to that product, not to the framework. Self-hosting means the operator supplies the surrounding controls.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.4 and v0.0.7, config_template.yamlread 2026-09-01

Cost is shown in dollars for every task, node and block, and runs stream live over WebSocket with step counts and activity summaries. Task history is searchable, and any entry opens in the builder with all execution steps preserved.

Per-model token consumption, cache hits and cost are tracked on every task, admin impersonation carries audit trails, and a read-only SQL Query block gives access to the customer's own platform data including custom analytics views. The execution diagnostics and analytics dashboards are admin tools for self hosters and are not exposed to cloud customers. No retention period for run history, or reasoning trace for AI blocks beyond inputs and outputs, is described.

Sourceagpt.co pricing page and block documentationread 2026-09-29

The Agent Performance Monitoring dashboard reports organization metrics for agents, runs and tokens, per-model breakdowns of runs and consumption, and agent-level analytics covering total runs, average tokens per run, total calls and run time.

Model, Tool and Knowledge Consoles each keep their own logs, so a finished run can be traced step by step through which model was called, which tools ran and which embeddings were queried.

A real-time activity feed shows execution as it happens, APM keeps data for deleted agents and runs, and the resource manager keeps a per-run output directory that downloads as a zip.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.8, v0.0.9 and v0.0.12read 2026-09-01

The platform is open source with a public repository and self-hosts free via Docker Compose, with the customer supplying their own model API keys so data and inference remain on their own infrastructure, and Llama among the pre-integrated models allows local model execution.

The hosted cloud provides Pro and Max subscriptions, with a Team plan marked coming soon, and a pay-as-you-go credit wallet. Platform code is licensed under Polyform Shield, which permits use, modification and self-hosting but bars building a competing product, while the wider repository is MIT. No region selection or data residency commitment for the hosted cloud is documented.

Sourcegithub.com/significant-gravitas/autogpt, agpt.co pricing page, docs.agpt.co self-hosting documentationread 2026-08-31

The framework is MIT-licensed and self-hosted with Docker Compose on the customer's own machine or infrastructure, and a separate GPU compose file supports local large language models and multi-GPU execution, so neither the platform nor the model has to leave the environment. A one-click Digital Ocean deployment runs the image on a self-hosted cloud machine and updates itself with each release. Storage can be local disk or S3, and the database and vector store belong to the operator. A hosted SuperAGI Cloud is also available for teams that prefer it.

Sourcegithub.com/TransformerOptimus/SuperAGI README installation section and release notes v0.0.4, v0.0.12 and v0.0.14read 2026-09-01

AutoGPT runs a public marketplace of community-built agents carrying ratings and real run counts, with one-click adds to a personal library. Library operation blocks add an agent from the store to a personal library using its store listing version ID, returning the library entry and agent graph IDs for execution, and retrieve all library agents with search and pagination, for uses such as agent provisioning, onboarding flows that seed starter agents, and adding agents on demand when a user asks for a capability. Store listing version IDs pin the adopted version.

Sourceagpt.co library and store operations block documentation, agpt.co home and pricing pagesread 2026-08-31

A browsable marketplace carries toolkits, agent templates, knowledge embeddings and models, reachable from the product and at its own subdomain. SuperCoder builds end-to-end coding projects from goals and instructions.

Predefined agent workflows include Sales Engagement, which fetches prospects from Apollo and drafts personalized outreach for each row, and Recruitment, which compares uploaded CVs against a job description and sends shortlist or rejection mail.

Customers can save templates from any run and publish them to the marketplace after a validated run, so the library grows from the community as well as from SuperAGI.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.6, v0.0.7, v0.0.10 and v0.0.12, marketplace.superagi.comread 2026-09-01

OpenAI, Anthropic, Groq and Llama models come pre-integrated, with the model chosen per block, and third-party integration guides show additional providers connected through AI blocks. Self-hosting is free with the customer supplying their own model API keys, while the hosted cloud uses a pay-as-you-go credit wallet for agent runs. AI reasoning blocks including AI Condition, Structured Response, AI Conversation, AI List Generator and Fact Checker each carry model configuration. No automatic routing between models at run time is described.

Sourceagpt.co pricing and home pages, docs.agpt.co platform documentationread 2026-08-31

Providers are configured in account settings, and a model is chosen per agent from a dropdown at provisioning. GPT-4-32k, gpt-3.5-turbo-16k and Google PaLM 2 Bison are supported, and a model factory class is built to admit more. HuggingFace and Replicate integrations let the customer bring any public or self-hosted model by entering the model name and inference endpoint, a models marketplace offers preconfigured entries, and local large language models run on the customer's own weights with multi-GPU support.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.7, v0.0.8, v0.0.11 and v0.0.14read 2026-09-01

A public REST API with an OpenAPI schema and generated SDK clients supports programmatic agent creation and execution, and a Block Development SDK lets customers build, register and ship custom blocks using a provider pattern that configures credentials and per run costs, with auto generated documentation surfacing custom blocks in builder search.

The platform is open source with a public repository and free self-hosting via Docker Compose; platform code is licensed under Polyform Shield with the wider repository MIT. The MCP tools block works in the other direction, with the platform calling out, and no first party MCP server exposes AutoGPT agents to external assistants.

Sourcegithub.com/significant-gravitas/autogpt, docs.agpt.co and dev-docs.agpt.co block SDK guide, agpt.co pricing pageread 2026-08-31

Public APIs, reachable from cloud or local base URLs, let developers create, pause, resume and update agents, start runs and retrieve run details and input or output files. API keys are generated in settings, and the full reference is published as a Postman collection. Python and NodeJS SDK client libraries install with pip or npm and have their own documentation section.

Webhooks send HTTP callbacks to a user's endpoint on five events, when an agent is running, paused, completed, terminated or reaches max iterations. Custom tools can be added by linking a GitHub repository from the frontend, with a cookie-cutter template and a community tools repository.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.11, v0.0.12 and v0.0.13, and documenter.getpostman.com SuperAGI collectionread 2026-09-01

The plan comparison lists Correctness and Accuracy Scoring, quality scores for each execution with alerts on regressions, and a dry-run self-repair loop in which newly generated workflows are simulated and errors caught and fixed automatically.

Run-from-builder execution, task presets and version-pinned schedules support testing a specific workflow version. Both appear only as single lines in that table, and the scoring method, test sets and recorded expected outputs are not described.

The execution analytics and diagnostics dashboards are admin tools for self hosters and are not exposed to cloud customers.

Sourceagpt.co pricing pageread 2026-09-29

The Coding toolkit ships a Write Unit Tests tool, and the SuperCoder sequence runs WriteSpec, WriteCode, ImproveCode and then WriteTest, so tests are written against the customer's generated project. SuperAGI does not say whether those tests are run, or whether a failing test blocks delivery or is reported back. Agent Performance Monitoring, the Model, Tool and Knowledge consoles and the activity feed report what happened after a run. There is no evaluation harness, scored comparison between agent versions or preview run against a held-out set.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.6 and v0.0.9read 2026-09-01

Stagehand Blocks bring AI powered browser automation through Stagehand and Browserbase: Stagehand Act performs click, fill, type, press, scroll and select actions described in natural language, alongside Extract and Observe, and each block requires the customer's Browserbase project ID. Browser automation also runs through a Chromium-backed agent browser, and self hosters bring their own keys and infrastructure for browser automation and other features that rely on third-party services. The browser engine is a third party service the customer connects, and the E2B sandbox block runs code.

Sourceagpt.co Stagehand blocks documentation and pricing pageread 2026-09-29

A Web Scraping tool in the toolkit set pulls data from webpages, alongside a Web Interaction toolkit. Both read the web through scraping and wired-in access and do not drive a real interface. SuperAGI has no hosted or local browser, desktop session or remote computer control, and its search tools are API-based.

Sourcegithub.com/TransformerOptimus/SuperAGI release notes v0.0.4 and README toolkit listread 2026-09-01