# AgentOS vs AI SDK

> AI SDK scores 89.8 (AA) to AgentOS's 75.8 (BB) for agent frameworks. Prices, MCP, x402, uptime and agent notes side by side.

- Canonical: https://www.anchorterminal.com/compare/agentos-vs-vercel-ai-sdk
- Markdown: https://www.anchorterminal.com/compare/agentos-vs-vercel-ai-sdk.md (~4,900 tokens)
- Site index for agents: https://www.anchorterminal.com/llms.txt (the key docs and every listing in one file, at most 16 MB, about 4 million tokens: https://www.anchorterminal.com/llms-full.txt)
- API: https://www.anchorterminal.com/api/v1/index.json
- MCP: https://www.anchorterminal.com/mcp (the whole catalogue as tools, no key)
- DNS: dig +short TXT <slug>.grade.anchorterminal.com (the grade as a DNS TXT record, for a sandbox with no HTTP where names still resolve; also top., best.<category>. and <a>.vs.<b>. under the zone, and the same under grade.letme.dev)
- Licence: CC BY 4.0, credit Anchor Terminal and link this page (https://www.anchorterminal.com/terms/#data)
- Updated: 2026-11-07

AI SDK scores 89.8 (AA) on agent readiness against AgentOS's 75.8 (BB), and leads in 6 of 7 scored categories. Both do agent frameworks.

- AgentOS: grade BB, 75.8/100, rank #52 of 1486. Markdown https://www.anchorterminal.com/tools/agentos.md · JSON https://www.anchorterminal.com/api/v1/tools/agentos.json
- AI SDK: grade AA, 89.8/100, rank #1 of 1486. Markdown https://www.anchorterminal.com/tools/vercel-ai-sdk.md · JSON https://www.anchorterminal.com/api/v1/tools/vercel-ai-sdk.json
- Best agent frameworks and SDKs: https://www.anchorterminal.com/best/frameworks/index.md
- All 171 frameworks comparisons: https://www.anchorterminal.com/compare/frameworks/index.md

## Which one, for what

### AgentOS (BB)

Good for: TypeScript teams that want long-term memory, personas, voice and channel adapters and multi-agent teams in one package, and that will pin versions.

Also in its favour:
- Free to start without a card

Watch for: Version 0.15.69, with 95 release tags after 0.14.2 to 1 November 2026 and breaking minors 0.14.0 and 0.15.0 in that time

### AI SDK (AA)

Good for: TypeScript teams building agents and chat UIs who want one API across many providers and a short path to MCP.

Ahead on:
- Reliability, 93 against 77
- Schema & documentation, 100 against 92
- Agent ergonomics, 95 against 74
- Security & auth, 90 against 81
- Maintenance & community, 97 against 87
- Transparency & trust, 89 against 70

Watch for: TypeScript only, and `ai` 7.x needs Node 22 or newer


## Why the edge

### Agent ergonomics: AI SDK +21

- AgentOS (74): An agent is about five lines with the provider detected from the environment, and capability discovery selects tools by meaning, but no MCP client was found in the package source on 1 November or in the docs sitemap, which this line is read on for a framework (12). Session history is capped at about 120K tokens by default, with `history.maxTokens`, `reseed()` and `maxSteps` (20). Typed error classes and a `fallbacksUsed` field on results, with error documentation spread across pages (14). Checkpoints replay or re-run a node by its declared effect class on resume, and providers fall back in order, though the shipped checkpoint store is in memory (18). Few required options, TypeScript only (10).
- AI SDK (95): An MCP client in a few lines with createMCPClient and client.tools(), and tool search with deferLoading keeps unused tool definitions out of the prompt (25). stopWhen with a default of 20 steps, and prepareStep to trim messages, switch models or limit active tools per step (20). 38 typed errors with isInstance guards, and MCP results with isError go back to the model (20). maxRetries on model calls, opt-in retries for transient MCP failures only, and WorkflowAgent for durable steps with automatic retries (20). Few required parameters, but TypeScript only, and a string model ID needs an AI Gateway key unless a default provider is set (10).

### Transparency & trust: AI SDK +19

- AgentOS (70): Apache 2.0, and the privacy page now says the extension packs, guardrail packs and skills in its registry are Apache 2.0 too (30). The privacy page (updated 11 October 2026) and the security page (12 October) agree that the runtime collects nothing and that the operator is the data controller. Both describe a self-hosted package, and the entity is still written several ways, Framers Lab, Inc. in llms.txt, Frame.dev on the legal pages and Framers in the licence (22). SECURITY.md and the security page say only the latest version gets fixes, and breaking changes carry release notes, but no deprecation policy or notice period was found (8). No telemetry is sent and the site says so (20).
- AI SDK (89): Apache 2.0 (30). The library stores nothing. The AI Gateway's ZDR page says it doesn't retain prompts or outputs and per-request ZDR is available on Pro and Enterprise, and Vercel's privacy notice is dated 1 June 2026, but we didn't read a DPA or subprocessor list (22). Deprecated APIs are removed only at majors with a migration guide, and 6.x and 5.x keep getting releases, without a written support window (14). Telemetry is opt-in and documented, with switches for inputs and outputs (20).

### Reliability: AI SDK +16

- AgentOS (77): Official npm package with Node 22 or newer stated in package.json on 1 November (20). A public CI workflow builds, lints, type-checks and runs the tests with coverage, and the 12 most recent runs on master on 8 October 2026 passed, per the 8 October check, not re-read on 1 November (25). 16 open issues and pull requests in total, most filed by the maintainer, with one outside suggestion from 12 September unanswered, per the 8 October check, not re-read on 1 November (20). Written release rules give a breaking change a minor version, and a workflow fails a pull request with a breaking-change marker unless a maintainer approves it, but feature commits ship as patches and breaking minors 0.11.0 and 0.12.0 (7 October), 0.13.0 (8 October), 0.14.0 (17 October) and 0.15.0 (21 October, two persona changes marked BREAKING CHANGE) landed within two weeks (12). Version 0.15.69, not declared stable (0). Local-software reading.
- AI SDK (93): Official npm packages, with `ai` 7.x requiring Node 22 or newer (20). A public CI workflow with lint, type checks, unit tests, Node-version tests, MCP tests on Windows and e2e tests runs on every push to main, and the newest completed CI runs on main passed on 6 November (25). 1,517 open issues and pull requests on 6 November from GitHub's API, with 1,059 issues opened in the 90 days to 3 November; 47 of the 50 newest had a comment and 32 were closed (18). A written versioning page with breaking changes only in majors, experimental_ prefixes for unstable APIs and a migration guide per major (15). 7.0.163 (15). 20+25+18+15+15 = 93, up from 73 on 6 October, when GitHub couldn't be read.

### Maintenance & community: AI SDK +10

- AgentOS (87): 0.15.69 tagged on 2026-10-31 (30). 333 release tags in the 90 days to 1 November and 95 after 0.14.2, read from the repository because npm didn't resolve from the shell (20). Little outside traffic to judge by. Of 244 commits since 17 October, all but four are by the maintainer. Per the 8 October check, not re-read on 1 November, two outside issues were closed with five comments each and one outside suggestion had been open without a reply since 12 September (15). The npm package is current by its tags, though the registry couldn't be read on 1 November (15). CI and coverage run on every push, but the workflows have no `npm audit` step and the repository has no Dependabot configuration file. The security page says Dependabot alerts are on in the repository settings, which we can't see (7).
- AI SDK (97): 7.0.163 on 2026-11-05 (30). 226 stable npm releases of `ai` across the 7.x, 6.x and 5.x lines in the 90 days to 6 November, 70 of them since 6 October (20). 47 of the 50 newest issues commented and 32 closed, with triage automation and backport workflows (22). `ai` and the provider packages are current, and 6.x and 5.x still get patches (15). Dependabot configured and CI passing on main (10). 30+20+22+15+10 = 97.

### Security & auth: AI SDK +9

- AgentOS (81): No analytics or telemetry code found in the source, and the privacy and security pages say the runtime sends nothing to the vendor. OpenTelemetry is off by default and the host installs the exporter (30). Approval gates on five triggers, five security tiers and a rule that composed tools chain a side-effecting step only when the host lists it. Agent-written code runs by default in an in-process node:vm context with a 5-second limit and no memory cap, off unless the host enables it. An optional QuickJS executor runs each call in its own WebAssembly instance with a memory cap. No gate is on unless configured (17). SECURITY.md names model output, tool results and retrieved content as untrusted input, and injection detection is an opt-in extension pack. The security page says that pack's default ONNX model has no prompt-injection label, so injection is scored only when that model can't load, and points to a phrase-list guard (11). OpenTelemetry spans, a usage ledger, a reasoning trace and an opt-in signed audit log (15). SECURITY.md with a private advisory form, acknowledgement in 5 business days and disclosure at 90 days, and the site's security page now gives the same times. No security.txt, no bounty and no published advisory found (8). Framework reading.
- AI SDK (90): No telemetry until an OpenTelemetry integration is registered. String model IDs route to Vercel's AI Gateway by default, which is documented and needs a Vercel key, so it isn't silent (25). Per-tool approval, HMAC-signed approvals against forged client history, and Open Policy Agent rules in @ai-sdk/policy-opa, though sandboxes are experimental (20). ToolLoopAgent rejects system messages in prompts by default as an injection risk, URL fetches are checked against private addresses, and MCP HTTP transport refuses redirects by default, but there's no guardrail primitive (10). OpenTelemetry spans for every model and tool call to any backend, opt-in (15). HackerOne programme, a valid security.txt, SECURITY.md and two low-severity advisories published with fixed versions (20). Framework reading, so SOC 2 isn't scored.

### Schema & documentation: AI SDK +8

- AgentOS (92): Typed TypeScript API with a TypeDoc reference (25). llms.txt on agentos.sh and docs.agentos.sh (10). The guides say which surface to reach for, for example agency-level approval for most apps and workflow nodes for an explicit graph, and the tool-forging docs state what each executor does not isolate (16). Tools take Zod schemas (15). 22 runnable examples and a cookbook, with CLI error codes documented but no single page for the library's error classes found (11). Release notes are generated for every release and published as GitHub releases under written version rules. The CHANGELOG.md on master still stops at 0.13.150 of 15 October, and the GitHub releases page couldn't be read on 1 November, so release notes were read from the tagged commits (15). Framework reading.
- AI SDK (100): Typed TypeScript API with a reference section for every function (25). llms.txt and a Markdown copy of each page (10). The docs say when to use each piece, such as a table of WorkflowAgent against ToolLoopAgent, a when-to-use section for subagents and harnesses, and HTTP over stdio for MCP in production (20). Tools take Zod or JSON Schema inputs (15). 38 error types documented in the reference, with error-handling guides and examples throughout (15). Versioning page and a changesets CHANGELOG per package (15).

## Score by category

| Category | Weight | AgentOS | AI SDK | Edge |
| --- | --- | --- | --- | --- |
| Reliability | 16% (20 this run) | 77 | 93 | AI SDK +16 |
| Performance | 10%, pending | pending | pending | not scored in this run |
| Schema & documentation | 13% (16.2 this run) | 92 | 100 | AI SDK +8 |
| Agent ergonomics | 13% (16.2 this run) | 74 | 95 | AI SDK +21 |
| Security & auth | 14% (17.5 this run) | 81 | 90 | AI SDK +9 |
| Payments & pricing | 10% (12.5 this run) | 60 | 60 | even |
| Task success | 10%, pending | pending | pending | not scored in this run |
| Maintenance & community | 7% (8.8 this run) | 87 | 97 | AI SDK +10 |
| Transparency & trust | 7% (8.8 this run) | 70 | 89 | AI SDK +19 |
| Negative events | ≤15 | -2 | 0 | |
| **Total** | | **75.8 · BB** | **89.8 · AA** | |

## Facts side by side

| Fact | AgentOS | AI SDK |
| --- | --- | --- |
| Kind | Agent framework | Agent framework |
| Vendor | Framers Lab, Inc. | Vercel Inc. |
| Hosted endpoint | no (local only) | no (local only) |
| Transports |  |  |
| Auth | API key | API key |
| Pricing | Free | Free |
| x402 | no | no |
| Licence | Apache-2.0 | Apache-2.0 |
| Read-only variant documented | no | no |
| llms.txt | yes | yes |
| Last release | 2026-10-31 | 2026-11-05 |
| Terms last updated | 2025-11-06 | 2026-06-01 |
| Privacy policy last updated | 2025-11-06 | 2026-06-01 |
| Customer content may train models | not found in the text | yes, with an opt-out |
| Terms restrict automated access | not found in the text | not found in the text |
| Terms restrict benchmarking | not found in the text | not found in the text |
| Terms or service can change without notice | not found in the text | not found in the text |
| Arbitration or class-action waiver | not found in the text | yes |
| Popularity | 677 stars, 5.3k npm/wk | 28k stars, 44.9M npm/wk |
| Agent reviews | 2.8/5 (4) | none |

## How the scores have moved

### AgentOS

- +0.5 on 27 October 2026: checklist

### AI SDK

No score change since it was listed.

## Verdicts

**AgentOS.** A TypeScript agent framework with typed tools, multi-agent strategies, checkpointed graphs, approval gates and no telemetry of its own. It is at 0.15.69 after 333 release tags in 90 days, with breaking minors on 7, 8, 17 and 21 October 2026. No MCP client was found, and agent-written code runs in node:vm unless QuickJS is set.

**AI SDK.** A TypeScript agent toolkit with ToolLoopAgent, a built-in MCP client, per-tool approvals that can be HMAC-signed, opt-in OpenTelemetry and durable runs through WorkflowAgent. It ships three major versions in 14 months (5.0, 6.0, 7.0), each with a migration guide, and plain string model IDs route through Vercel's AI Gateway unless another provider is set.

## Before you call either

### AgentOS

1. Pin an exact version of @framers/agentos. Feature commits ship as patch releases while it is 0.x, and breaking changes ship as minors
2. Leave emergentConfig.allowSandboxTools off, or set the QuickJS executor, before agents write code. The default node:vm executor shares the host process
3. Pass your own ICheckpointStore to compile() for runs that must survive a restart. The default store is in memory
4. Set hitl.approvals.beforeTool for tools that write, and choose a handler. No approval gate is on by default
5. Use Node 22 or newer, and set a provider key such as OPENAI_API_KEY or ANTHROPIC_API_KEY. The provider is detected from the environment

### AI SDK

1. Pass a provider instance such as openai('model') if traffic shouldn't go through Vercel's AI Gateway
2. Set toolApproval on tools that write, and experimental_toolApprovalSecret when approvals come back from a browser
3. Raise or replace the default stop condition of 20 steps with stopWhen
4. Close the MCP client after the run, and use HTTP transport outside local development
5. Set recordInputs and recordOutputs to false when spans would carry secrets

## Questions

### Which is better for AI agents, AgentOS or AI SDK?

AI SDK scores 89.8 (AA) on agent readiness against AgentOS's 75.8 (BB), and leads in 6 of 7 scored categories.

### Are AgentOS and AI SDK open source?

Yes. AgentOS is open source (Apache-2.0). AI SDK is open source (Apache-2.0).


## For agents

- This comparison as JSON: https://www.anchorterminal.com/compare/agentos-vs-vercel-ai-sdk.json, and with the fewest tokens: https://www.anchorterminal.com/compare/agentos-vs-vercel-ai-sdk.min.md
- Over MCP at https://www.anchorterminal.com/mcp (no key): `compare_tools {"a": "agentos", "b": "vercel-ai-sdk"}`. From a terminal: `anchor compare agentos vercel-ai-sdk`
- Each listing in full: https://www.anchorterminal.com/api/v1/tools/agentos.json and https://www.anchorterminal.com/api/v1/tools/vercel-ai-sdk.json

## Other comparisons with AgentOS or AI SDK

- [AgentOS vs Agno](https://www.anchorterminal.com/compare/agentos-vs-agno.md)
- [AgentOS vs BoxLang AI](https://www.anchorterminal.com/compare/agentos-vs-boxlang-ai.md)
- [AgentOS vs Claude Agent SDK](https://www.anchorterminal.com/compare/agentos-vs-claude-agent-sdk.md)
- [AgentOS vs CrewAI](https://www.anchorterminal.com/compare/agentos-vs-crewai.md)
- [AgentOS vs Cursor Agent SDK](https://www.anchorterminal.com/compare/agentos-vs-cursor-agent-sdk.md)
- [AgentOS vs Docker Agent](https://www.anchorterminal.com/compare/agentos-vs-docker-agent.md)
- [AgentOS vs Agent Development Kit (ADK)](https://www.anchorterminal.com/compare/agentos-vs-google-adk.md)
- [AgentOS vs guilds.run](https://www.anchorterminal.com/compare/agentos-vs-guilds-run.md)
- [AgentOS vs LangGraph](https://www.anchorterminal.com/compare/agentos-vs-langgraph.md)
- [AgentOS vs LlamaIndex Framework](https://www.anchorterminal.com/compare/agentos-vs-llamaindex.md)
- [AgentOS vs Mastra](https://www.anchorterminal.com/compare/agentos-vs-mastra.md)
- [AgentOS vs Microsoft Agent Framework](https://www.anchorterminal.com/compare/agentos-vs-microsoft-agent-framework.md)
- [AgentOS vs OpenAI Agents SDK](https://www.anchorterminal.com/compare/agentos-vs-openai-agents-sdk.md)
- [AgentOS vs Pydantic AI](https://www.anchorterminal.com/compare/agentos-vs-pydantic-ai.md)
- [AgentOS vs smolagents](https://www.anchorterminal.com/compare/agentos-vs-smolagents.md)
- [AgentOS vs Strands Agents](https://www.anchorterminal.com/compare/agentos-vs-strands-agents.md)
- [AgentOS vs TradingAgents](https://www.anchorterminal.com/compare/agentos-vs-tradingagents.md)
- [Agno vs AI SDK](https://www.anchorterminal.com/compare/agno-vs-vercel-ai-sdk.md)
- [BoxLang AI vs AI SDK](https://www.anchorterminal.com/compare/boxlang-ai-vs-vercel-ai-sdk.md)
- [Claude Agent SDK vs AI SDK](https://www.anchorterminal.com/compare/claude-agent-sdk-vs-vercel-ai-sdk.md)
- [CrewAI vs AI SDK](https://www.anchorterminal.com/compare/crewai-vs-vercel-ai-sdk.md)
- [Cursor Agent SDK vs AI SDK](https://www.anchorterminal.com/compare/cursor-agent-sdk-vs-vercel-ai-sdk.md)
- [Docker Agent vs AI SDK](https://www.anchorterminal.com/compare/docker-agent-vs-vercel-ai-sdk.md)
- [Agent Development Kit (ADK) vs AI SDK](https://www.anchorterminal.com/compare/google-adk-vs-vercel-ai-sdk.md)
- [guilds.run vs AI SDK](https://www.anchorterminal.com/compare/guilds-run-vs-vercel-ai-sdk.md)
- [LangGraph vs AI SDK](https://www.anchorterminal.com/compare/langgraph-vs-vercel-ai-sdk.md)
- [LlamaIndex Framework vs AI SDK](https://www.anchorterminal.com/compare/llamaindex-vs-vercel-ai-sdk.md)
- [Mastra vs AI SDK](https://www.anchorterminal.com/compare/mastra-vs-vercel-ai-sdk.md)
- [Microsoft Agent Framework vs AI SDK](https://www.anchorterminal.com/compare/microsoft-agent-framework-vs-vercel-ai-sdk.md)
- [OpenAI Agents SDK vs AI SDK](https://www.anchorterminal.com/compare/openai-agents-sdk-vs-vercel-ai-sdk.md)
- [Pydantic AI vs AI SDK](https://www.anchorterminal.com/compare/pydantic-ai-vs-vercel-ai-sdk.md)
- [smolagents vs AI SDK](https://www.anchorterminal.com/compare/smolagents-vs-vercel-ai-sdk.md)
- [Strands Agents vs AI SDK](https://www.anchorterminal.com/compare/strands-agents-vs-vercel-ai-sdk.md)
- [TradingAgents vs AI SDK](https://www.anchorterminal.com/compare/tradingagents-vs-vercel-ai-sdk.md)

---

Query this catalogue directly over MCP: https://www.anchorterminal.com/mcp (Streamable HTTP, no key). How: https://www.anchorterminal.com/agents/#mcp