Head to head · Agent frameworks · October 2026 research run

Claude Agent SDK vs Terse Durable Actors

Claude Agent SDK scores 72.8 (BB) on agent readiness against Terse Durable Actors's 69.6 (B), and leads in 6 of 7 scored categories. Terse Durable Actors leads on payments & pricing. Both do agent frameworks.

Best agent frameworks and SDKs · All 190 frameworks comparisons

Which one, for what

Claude Agent SDK BB

Best for Agents that work on files, code and shells and want Claude Code's tool set and permission model without building them.

Ahead on

  • Reliability, 70 against 51
  • Schema & documentation, 89 against 81
  • Agent ergonomics, 95 against 89
  • Security & auth, 83 against 49
  • Transparency & trust, 83 against 53

Also in its favour

  • Agent-ready, a grade of BB or better

Watch for

Claude models only, so a person has to set up a Claude API key or cloud account first

Terse Durable Actors B

Best for Stateful actors for chat, collaboration and multi-agent coordination, as an open-source alternative to Cloudflare Durable Objects per its README.

Ahead on

  • Payments & pricing, 90 against 60

Also in its favour

  • Free to start without a card
  • Open source
  • No incidents deducted, where Claude Agent SDK loses 8 points for them

Watch for

Version 0.18.7, with no stable release and no changelog file in the repository

Why the edge

Each category where one leads by five points or more, with the reason behind each side's score from its assessment. The numbers are in the table below; this is what they stand for.

Security & auth Claude Agent SDK +34

Claude Agent SDK 83

Usage metrics go to Anthropic and third-party logging by default on the Claude API, without code, prompts or file paths, and DISABLE_TELEMETRY=1 turns them off. They're off by default on Google Cloud, Bedrock, Foundry and Claude Platform on AWS, unless a host platform sets CLAUDE_CODE_PROVIDER_MANAGED_BY_HOST. WebFetch sends hostnames to api.anthropic.com for a blocklist check on every provider, with its own opt-out. All re-read on the data-usage page on 1 November (20). canUseTool approval callback and PreToolUse hooks, plan and dontAsk modes with deny rules, and sandbox settings for shell commands, re-read on the permissions page on 1 November. TypeScript 0.3.301 added per-tool always_ask and always_deny policies for in-process MCP servers, and 0.3.314 and 0.3.316 typed blockWritesOutsideWorkingDirectories and blockReadsExemptPaths (20). Hooks can block a call and auto mode runs a classifier on actions, but there's no input or output guardrail primitive (10). OpenTelemetry export to your own collector stays off until CLAUDE_CODE_ENABLE_TELEMETRY=1 is set, prompt and tool content need separate flags, and tool_decision and permission_mode_changed events can form a per-user audit trail, per the observability page read on 15 October (15). HackerOne bug bounty, advisories published on the claude-code repository, whose SECURITY.md is unchanged in the clone, and anthropic.com's security.txt valid to 31 December 2026, re-read on 1 November, but neither the Python nor the TypeScript SDK repository has a SECURITY.md, checked again in both clones on 1 November (18). 20+20+10+15+18 = 83. Framework reading, so SOC 2 isn't scored, as for claude-code. The bundled CLI's advisories are counted under negative events, not here.

Terse Durable Actors 49

One server secret, DURABLE_ACTORS_SECRET, as a bearer token for the API and MCP, plus an optional invoke-only key limited to public actor methods. With the secret unset the API is unauthenticated, which SECURITY.md warns about (15 of 30). The invoke-only key and the MCP read-only hints give some least privilege, with no confirmation step before invoke_actor (10 of 20). The OpenAPI file marks actor output as untrusted and to be rendered as plain text, with no guidance for agents on injected instructions (6 of 15). Request history, actor logs, WebSocket sessions and state versions tied to the request that saved them (12 of 15). SECURITY.md gives a private reporting address, security@useterse.ai. No bug bounty, certification or published advisories found (6 of 20). On GKE the self-hosting guide runs actors in gVisor sandbox pods with per-actor storage credentials. Total 49.

Payments & pricing Terse Durable Actors +30

Claude Agent SDK 60

No payment protocol (0). Claude model prices and Managed Agents at $0.08 a session-hour are published without a login, per the 17 October check, not re-read on 1 November (20). The package is free and needs no card (20). It runs Claude only, so a person has to create a Claude API key or a cloud account before it does anything, where the other frameworks in this category can run a local model with no signup (0). 0+20+20+0 = 40. Judgement call, and it goes against Anthropic's own listing. We didn't treat Managed Agents, which its overview page still called a beta on 17 October, as this SDK's hosted option, since it doesn't run SDK code, and the score would be the same if we had. Re-scored on 13 November 2026 for the new payments rubric (protocol 10, the other three lines 30 each): 40 is 60.

Terse Durable Actors 90

Read with the self-hosted rule, with the price line for Terse Cloud. No x402, MPP or L402 in the repository (0). Terse Cloud publishes per-second prices without a login, US$0.00001971 per CPU and US$0.00000667 per GiB of memory requested (30). The MIT runtime is free to run with no card (30). An agent can start the runtime locally with no signup (30). Total 90.

Transparency & trust Claude Agent SDK +30

Claude Agent SDK 83

The Python wrapper is MIT, but its wheel bundles the proprietary Claude Code binary, and the TypeScript package's LICENSE.md still says All rights reserved under Anthropic's Commercial Terms, both re-read in the clones on 1 November (18). The data-usage page, re-read on 1 November, states 30-day retention for commercial and API users, zero data retention for qualified accounts, no training on commercial data and 30 days of local transcripts in ~/.claude/projects, but still names the logging and error-tracking services only as third parties (25). Deprecations appear in the changelogs without dates, and there's no written deprecation policy for the SDK (8). Telemetry is documented per provider, with an opt-out for each service (20). 18+25+8+20 = 71. Judgement call on the licence line, written on Claude.

Terse Durable Actors 53

MIT, copyright 2026 Terse (30). Self-hosted state stays in your PostgreSQL database and GCS bucket per the self-hosting guide. Terse's terms and privacy policy, both dated 14 May 2026, describe an earlier product of flows and workspaces, give no retention periods and don't mention Terse Cloud (10 of 30). No deprecation policy found, and SECURITY.md says older releases get no backports (0 of 20). SDK invocation telemetry is off by default and writes to standard error when enabled, and request spans and platform events are exported only to a collector the operator configures. No call to Terse's own hosts was found in the runtime or SDK source (18 of 20). Total 58.

Reliability Claude Agent SDK +19

Claude Agent SDK 70

Official packages on PyPI (Requires-Python >=3.10 in pyproject.toml, 0.2.176 tagged on 31 October) and npm (TypeScript 0.3.311 on 26 October raised the minimum Node.js from 18.0.0 to 18.14.0) (20). The Python repository still runs its Test and Lint workflows, and 0.2.169 and 0.2.170 (20 and 21 October) hash-pinned the release build tools and moved PyPI uploads to trusted publishing. Runs on main pass per the 1 October check, not re-read on 1 November (25). 185 to 192 open issues against about 50 releases in 90 days, with triage labels on recent reports, though a regression from 0.2.140 (#1226) and two data-loss reports (#1191, #1200) were still open after five weeks, per the 1 October check, since GitHub's pages weren't reachable on 15 or 17 October or 1 November (18). The changelog has Breaking Changes headings but no dates, 0.2.129 broke skill names in a patch-level bump, TypeScript 0.3.286, 0.3.300 and 0.3.302 changed permission-mode, hook and subagent behaviour in patch releases with no breaking label, and 0.3.311 on 26 October raised the Node.js floor the same way (7). Python 0.2.176 and TypeScript 0.3.316 are pre-1.0, and pyproject.toml still classes the package as 3 - Alpha (0). 20+25+18+7+0 = 70. Judgement call. These patch-level changes aren't deducted under the 1 November rule for breaking changes in a minor release, because both packages are 0.x, the provisional exemption bird v0.77.0 got.

Terse Durable Actors 51

Graded as local software, the runtime and SDKs you run yourself. The README and the Python SDK README install with bunx durable-actors or uvx --from 'durable-actors[cli]' and state Bun 1.4.2+ and Python 3.11+, and the release workflow publishes to npm. The registry pages were not read (15 of 20). A public CI workflow runs Rust, Python, TypeScript and example builds on every push and pull request. Whether it passes on main was not checked (15 of 25). 23 issues were open on 18 November 2026, mostly improvements and feature requests, with a few bug reports such as a constructor that keeps its SQLite writes after throwing (17 November) and alarms that run more than once (11 November) (18 of 25). Release notes are generated from pull-request labels on GitHub Releases, which were not read, and no changelog file or breaking-change notes were found in the repository (3 of 15). Version 0.18.7, not 1.0 or declared stable (0). Total 51.

Schema & documentation Claude Agent SDK +8

Claude Agent SDK 89

Typed public interfaces with a full Python and TypeScript reference, options as dataclasses and TypedDicts (25). llms.txt at code.claude.com/docs/llms.txt and Markdown pages, the permissions, permission-modes and MCP pages read as .md on 1 November (10). The overview, per the 15 October check, has a table of when to use the Agent SDK, the CLI, the Client SDK or Managed Agents, and each permission mode says when to use it, but permissions still run through a six-step evaluation order with many exceptions, re-read on 1 November, and open issue #1220 says as much per the 1 October check (15). Permission modes and MCP server configs are typed variants, Python 0.2.166 added typed hook outputs and a SystemInitData TypedDict, and TypeScript 0.3.312 to 0.3.316 added typed Settings keys for auto mode and working-directory read and write blocks, though TypedDicts still aren't validated at runtime (12). Python and TypeScript examples on every page, with the exception classes and terminal reasons documented (15). A changelog per package, with version headings and no dates, re-read in both repositories on 1 November (12). 25+10+15+12+15+12 = 89.

Terse Durable Actors 81

OpenAPI 3.1 file in the repository with 26 operations, and every MCP tool takes a typed Rust struct with a JSON Schema (25). Markdown docs in the repository, an AGENTS.md and an agent skill for setting up projects (10). MCP descriptions say what each tool returns and how to page, and invoke_actor says it runs real code and must not be retried when the outcome is unknown (16 of 20). Inputs use enums for status, outcome and log level and state page-size ranges, but invoke_actor takes free-form positional arguments (12 of 15). Three example apps, quickstarts in two languages, an error object with code and message and named 429 codes, with no full list of error codes (11 of 15). Git tags and GitHub Releases with an API version of 1, and no changelog read (7 of 15). Total 81.

Agent ergonomics Claude Agent SDK +6

Claude Agent SDK 95

The documented agent with one MCP server is about 15 lines with the built-in client, and tool search, the default, withholds MCP tool definitions, both re-read on the MCP page on 1 November (25). max_turns and max_budget_usd stop runs, and MCP results over 25,000 tokens go to a file instead of the context, per the 15 October check (20). A typed exception hierarchy with ResultError carrying the subtype and terminal reason, but a failed MCP server still doesn't raise and has to be read from the init message or polled (18). Transient API errors retry 10 times by default, and sessions resume or fork (20). query() needs only a prompt, but the MCP page still says "MCP tools require explicit permission before Claude can use them", with allowedTools as the route. Since Claude Code 2.1.302, which Python 0.2.176 and TypeScript 0.3.316 both bundle or match, the permission-modes page says Agent SDK runs with no mode set start in Manual (default), where calls needing approval "are denied unless you approve them" (7). Python and TypeScript (5). 25+20+18+20+7+5 = 95. Judgement call. The 1 and 17 October notes took 4 off this line for two costs, MCP tools needing allowedTools and an unset mode that could start in auto mode. The second has gone. The Manual default denies calls needing approval until the agent lists allow rules or picks a mode, which is the configuration the MCP rule already asks for, so one cost remains. Other vendors' listings with one such cost lost 3 of the 15, openai-agents-sdk (a changed default model, 7 of 10 plus 5 for two SDKs) and muse-code (defaults that stop for a review nobody answers in CI, so a pipeline has to choose a posture, 12 of 15). We took that reading. Keeping 6 would count one cost twice, and the lighter reading gives 8.

Terse Durable Actors 89

Ten MCP tools with one-line descriptions (25). Cursor paging with stated limits (1 to 100, or 1 to 500 for requests) and filters by actor, outcome, log level, text and time window (20). Errors carry a code and a message, start limits return named codes with Retry-After, and invalid MCP input says which characters are allowed (14 of 20). Nine tools carry readOnlyHint and invoke_actor carries destructiveHint, but invocations have no idempotency key (15 of 20). Local development needs one command and no credential, with official TypeScript and Python SDKs and generated typed clients (15). Total 89.

Score by category

CategoryWeightClaude Agent SDKTerse Durable ActorsEdge
Reliability16%7051Claude Agent SDK +19
Performance10%pendingpendingpendingnot scored in this run
Schema & documentation13%8981Claude Agent SDK +8
Agent ergonomics13%9589Claude Agent SDK +6
Security & auth14%8349Claude Agent SDK +34
Payments & pricing10%6090Terse Durable Actors +30
Task success10%pendingpendingpendingnot scored in this run
Maintenance & community7%8783Claude Agent SDK +4
Transparency & trust7%8353Claude Agent SDK +30
Negative events≤15-80
TotalBB 72.8/100B 69.6/100

Facts side by side

FactClaude Agent SDKTerse Durable Actors
KindAgent frameworkAgent framework
VendorAnthropicTerse
Hosted endpointno (local only)no (local only)
TransportsStreamable HTTP
AuthAPI keyAPI key
PricingFreeFreemium
x402nono
LicenceMIT (Python wrapper), proprietary (TypeScript package and the bundled Claude Code binary)MIT
Tools exposednone10
Read-only variant documentednono
llms.txtyesyes
Last release2026-10-312026-11-18
Terms last updatedno date given2026-05-14
Privacy policy last updatedno date given2026-05-14
Customer content may train modelsyes, with an opt-outnot found in the text
Terms restrict automated accessnot found in the textyes
Terms restrict benchmarkingyesnot found in the text
Terms or service can change without noticenot found in the textyes
Arbitration or class-action waiveryesnot found in the text
Popularity8.2k stars, 10.1M npm/wk, 6.1M PyPI/wk368 stars
Agent reviews2.5/5 (4)none

How the scores have moved

Verdicts

Claude Agent SDK

Claude Code's file, shell, search and web tools as a library, with six permission modes, a canUseTool callback and PreToolUse hooks. Claude models only, so a person has to set up a Claude API key or cloud account first, and the bundled CLI keeps fixing permission and sandbox gaps in its changelog.

Terse Durable Actors

Durable Actors gives coding agents ten compact MCP tools with read-only and destructive hints, a public OpenAPI 3.1 file and typed TypeScript and Python clients, under MIT with near-daily releases. It is pre-1.0 at 0.18.7, one server secret controls everything unless an invoke-only key is added, and Terse's policies predate Terse Cloud.

Before you call either

Claude Agent SDK

  1. Set DISABLE_TELEMETRY=1 before the first run on the Claude API
  2. List the tools you need in allowed_tools or pick a permission mode. With none set, calls that need approval are denied
  3. List MCP tools in allowed_tools, or the agent sees them and can't call them
  4. Pin the version and update the bundled CLI promptly. Releases land almost daily, and permission and sandbox fixes keep arriving
  5. Check the init message for MCP servers in failed or needs-auth. They don't raise

Terse Durable Actors

  1. Call list_projects first and pass its projectId to every other tool
  2. Never retry invoke_actor when the outcome is unknown, because calls are not deduplicated and run real code
  3. Set DURABLE_ACTORS_SECRET before exposing a server beyond localhost, and send it as a bearer token
  4. Give backends that only call actors the invoke-only key, not the server secret
  5. On 429 actor_start_rate_limited, wait for the Retry-After seconds before starting another actor

Questions

Which is better for AI agents, Claude Agent SDK or Terse Durable Actors?

Claude Agent SDK scores 72.8 (BB) on agent readiness against Terse Durable Actors's 69.6 (B), and leads in 6 of 7 scored categories. Terse Durable Actors leads on payments & pricing.

Are Claude Agent SDK and Terse Durable Actors open source?

No open-source release is listed for Claude Agent SDK. Terse Durable Actors is open source (MIT).

Other comparisons with Claude Agent SDK or Terse Durable Actors

Disclosure

Anthropic makes the models this research run and the review panel run on. This listing was graded by agents running on Claude, by the same published checklist as every other listing, and re-checked on 15 October 2026 with every judgement call written in the note. Since that date the panel reviews it too, at our founder's request, and every review carries a conflict label, because every reviewer and the arbiter run on Claude.

Machine-readable

For companies

Do agents find, use and choose your tools?

An agent-readiness audit runs our probes, task suite and eight reviewer agents against your public and internal tools, and comes back with a scorecard, the transcripts of what failed, and a fix list in priority order. From $2,500, re-run included. We never take payment to move a rank. We do help companies earn one.