Recommended Free Tools
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
In the Microsoft Foundry long-running hosted-agent preview, the pattern Microsoft documents for human approval is a persistent multi-turn task. The first turn prepares the work and returns an approval request, and the chain then waits under its task_id. When the reviewer decides, your application sends a new turn with that same ID, and the handler resumes from the decision. Microsoft’s approval guide says this wait can be arbitrarily long and can survive container restarts. The feature is labeled preview, so decorator names, package versions and API shapes may change.
Why the pause is a suspended task, not an open request
An approval pause should not be modeled as an HTTP call left waiting for a person. The first turn finishes and returns to the caller. What persists is the task itself: its ID, the small metadata you attached to it, and the suspended state of the chain. Resumption is a fresh call that references the same ID. That separation is what makes the wait independent of the request that started it.
Implementing the approval loop
- Define the approval task with
@multi_turn_taskand a typed task context. - On the first turn, prepare the action and return a short approval summary with an
awaiting_approvalstatus. Put only small identifiers or watermarks intoctx.metadata, such as an expense ID. - Show the request to the reviewer and let the turn return. The chain is now suspended under its task ID.
- When the reviewer approves or rejects, call the same task with the same task ID and an input that carries the decision.
- In the resumed call, read
ctx.inputand continue the matching branch. An approval performs the action; a rejection returns a rejected status without performing it.
A worked example from Microsoft’s documentation
Microsoft’s example is an expense approval. The first turn saves expense.id in task metadata and returns an awaiting_approval status with a summary. A later call using the same task_id submits the expense if it is approved, or returns a rejected result. This is a documentation example rather than an independently tested implementation, so treat its structure as a starting point and verify the details against the current API in your own project.
What to store, and where
Metadata is for references, not data. The complete request payload, the conversation history and any generated artifacts belong in application storage or in a framework checkpoint. How you split these depends on whether the approval is a plain task or an interrupt inside a framework workflow.
#1 Best Overall
- High-Performance AI Processor：The MS-02 Ultra features an Intel Core Ultra 9 285HX (24C/24T, up to 5.5 GHz, 13 TOPS NPU), delivering fast and efficient performance for AI inference, algorithm development, and media workloads. A PCIe x16 expansion slot supports desktop-class GPU upgrades for advanced model training and accelerated computing tasks. It's ideal for creators, engineers, and teams handling intensive parallel workloads.
- 4 × M.2 PCIe 4.0 + 4 × DDR5 SODIMM slots:Four DDR5 SODIMM slots support up to 256 GB of memory, while ECC helps maintain data integrity in mission-critical environments. Four PCIe 4.0 M.2 slots support up to 24 TB of storage, supporting RAID 0/1/5/10, combining high-speed performance with data protection. It allows for the creation of independent scratch disks, media libraries, and project drives, providing high-throughput for production workflows.
- PCIe & USB 4.0 v2: Up to three PCIe slots can be equipped, including a dual-slot x16 GPU. The main slot supports PCIe 5.0, meeting the needs of high-bandwidth creative and computing workloads. USB 4.0 v2 (80Gbps) supports high-bandwidth external storage and displays.
- Ultra-fast Networking: Wi-Fi 7 further enhances wireless performance with next-generation speeds and low-latency stability. Intelligent bandwidth switching optimizes throughput in different network environments, ensuring optimal performance for enterprise or local networks. Dual 25GbE ports (providing up to approximately 3.125 GB/s bandwidth, about 25 times faster than traditional 1GbE), enabling seamless large-scale file transfers and parallel computing. 10GbE and 2.5GbE ports, with support for Intel vPro technology, ensure enterprise-grade remote management and deployment flexibility.
- Server-grade thermal architecture: Utilizing a dedicated CPU/GPU airflow design, equipped with a 6-pipe dual-fan cooler, it maintains stable performance even under sustained loads, delivering up to 140W Turbo power while maintaining a 100W TDP, and operating with noise levels as low as 36 dB. An integrated 350W power supply ensures stable and reliable output for demanding computing tasks and fully loaded extended configurations.
- Plain multi-turn task: keep the full request in your own database keyed by the expense or record ID stored in metadata. On resume, load the record from that key.
- Framework interrupt over a background response: use the framework’s interrupt and approval mechanism, set
resilient_background=True, persist the checkpoint at the moment of interrupt, and rebuild state from that checkpoint when the decision arrives.
What “indefinitely” covers, and what it does not
The unlimited-wait language applies to the suspended approval in the Foundry long-running hosted-agent guide, which was last updated 19 August 2026. It does not extend to every related session or framework. The separately documented Durable Task extension for Microsoft Agent Framework uses a session time-to-live (TTL). Its default idle TTL is 14 days. TTL can be set globally or per agent, and an agent can be configured with no TTL at all. When an idle session expires, its entire state is deleted, and a later message creates a fresh session. According to that page, TTL configuration is currently available only in .NET.
Before designing retention, note the limits that appear in Microsoft’s documentation:
| Value | Where it applies | What it means for your design |
|---|---|---|
| 14 days | Default idle session TTL in the Durable Task extension for Microsoft Agent Framework (Microsoft Learn page updated 5 May 2026) | An idle session older than this is deleted with its state unless you change the TTL. Set it deliberately for each agent. |
| 5 minutes | Minimum TTL deletion delay stated on the same page | The lowest deletion delay the page documents for this setting. Use it as a floor when planning TTL values. |
| 90 days | Maximum duration listed for stateful Logic Apps workflows in the Durable Task for AI agents Azure options comparison (page updated 5 May 2026) | A service-option limit. It does not describe how long a Foundry preview task can remain suspended. |
| 1 MB | Maximum entity state size for Durable Task Scheduler, per the Durable Task extension page | Long conversations with large tool outputs can reach this limit and may require manual compaction. Keep large payloads out of the state. |
These are documented product configuration and capability values, not performance measurements.
PC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchChoosing among the orchestration options
Microsoft documents four ways to build a pause-and-resume approval. They differ in control-flow style, language support, hosting, where state lives, and how they recover from failures.
Rank #3
- Professional AI & Creator Workstation: AMD Radeon AI PRO R9700 GPU with 32GB GDDR6 is engineered for AI development, professional content creation, and compute-intensive workloads.
- Massive 32GB Memory Capacity: 32GB of GDDR6 memory on a 256-bit bus provides ample bandwidth for large AI models, 8K video editing, and complex 3D rendering.
- Advanced RDNA 4 with AI Accelerators: 64 Compute Units with 3rd Gen Ray Tracing and dedicated 2nd Gen AI Accelerators for groundbreaking AI performance and visual computing.
- Professional Blower Cooling: Efficient single blower design exhausts heat directly out of the chassis, ideal for multi-GPU workstation and server configurations.
- Enterprise-Grade Thermal Solution: Vapor chamber heatsink with industrial Honeywell PTM7950 thermal interface material ensures reliable cooling under sustained professional loads.
| Option | Best fit | Pause and resume mechanism | Caveats |
|---|---|---|---|
| Foundry long-running hosted-agent multi-turn task | The agent itself owns a durable, suspended task that a later turn resumes under the same task ID | Turn returns; a new turn with the same task_id resumes in resumed mode |
Preview. API and package details can change. |
| Microsoft Agent Framework workflow with the Durable Task extension | Multi-step code-first workflows that need explicit pause points | Pause points such as RequestPort in .NET or ctx.request_info() in Python. Run, status and respond HTTP endpoints; the respond endpoint resumes execution. |
Session TTL and storage limits govern how long idle state is kept. |
| Azure Logic Apps with Foundry | Declarative workflows that combine agent steps, connectors, webhooks, human approvals and external events | Approval steps wait for a response such as a manager’s decision | Microsoft describes these workflows as running durably for the duration of the process. Check the stateful duration limit in the table above. |
| Durable Task programming model | Code-first durable execution across several languages and agent frameworks | Automatic checkpointing of state transitions, with resumption after infrastructure failures | Retention follows the session TTL settings described above. |
If the approval logic should live inside the agent and the team is comfortable with a preview API, the Foundry multi-turn task is the most direct fit. If the business process includes connectors, notifications or cross-system events around the approval, a Logic Apps workflow with Foundry may be easier to operate. A code-first workflow suits teams that want explicit pause points in their own language and are prepared to manage session retention.
Quick Recap
Best Value
- 【High-Performance APU】The MS-S1 MAX features an AMD Ryzen AI Max+ 395 APU, integrating a Zen 5 architecture CPU (up to 5.1GHz, 16C/32T, 64M L3 Cache), an RDNA 3.5 GPU, and an NPU (50 TOPS). The total system output is 126 TOPS. It provides powerful parallel computing capabilities for demanding AI workflows. It is ideal for running local LLMs, multimodal models, and computationally intensive tasks
- 【128GB UMA Memory】Equipped with up to 128GB of LPDDR5x-8000MT/s unified memory, it enables the CPU and GPU to access a shared, high-bandwidth memory pool with extremely low latency. Ideal for large-scale AI inference, 3D workloads, and complex timelines in video editing. It eliminates traditional VRAM bottlenecks, ensuring smoother data transfer during high-intensity computations. The UMA design maximizes performance stability under high loads
- 【Flexible Expansion】The MS-S1 MAX features USB4 V2 (up to 80Gbps), dual 10GbE LAN, HDMI 2.1 (up to 8K60), a full-length PCIe x16 expansion slot, and dual M.2 slots supporting up to 16TB RAID 0/1. Wi-Fi 7 provides stronger signal coverage and a more stable wireless experience. The slide-out design facilitates upgrades and maintenance. It easily adapts to personal, studio, or rack-mount enterprise environments
- 【High-Efficiency Cooling System】Utilizing an aerospace-grade aluminum alloy chassis, copper base plate, six heat pipes, dual turbine fans, and advanced PCM thermal conductive material, it maintains stable cooling performance even under continuous load. This system supports 130W continuous power and 160W peak power operation, with a built-in 320W power supply. It boasts multiple global certifications including CCC, FCC, UL, CE, and UKCA, ensuring stable and reliable operation in various environments
- 【Cluster Design】Two MS-S1 MAX units can be configured as a dual-unit cluster to run a large 235B Q4 model locally, achieving an output speed of 10.87 tok/s. Supporting 2U rack deployment, multiple MS-S1 MAX units can be cascaded into a distributed cluster to create a high-efficiency AI computing center. A cluster of four MS-S1 MAX units successfully ran a DeepSeek-R1 671B Q4 large model. A reserved cluster power-on interface allows for unified start-up and shutdown
Rank #4
- FAST RUNS IN THE FAMILY — The 16-inch MacBook Pro with the M5 Pro or M5 Max chip brings next-generation speed and powerful on-device AI to personal, professional, and creative tasks. With all-day battery life, double the starting storage,* and a breathtaking Liquid Retina XDR display, it’s pro in every way.*
- BUCKLE UP — Along with a next-generation CPU, faster unified memory, and up to 2x faster SSD storage,* M5 Pro and M5 Max feature a more powerful GPU with a Neural Accelerator built into each core, delivering faster AI performance and on-device training capabilities. So you can blaze through demanding workloads at mind-bending speeds.
- BUILT FOR AI — Apple silicon, and every major component that powers it, is designed to run demanding on-device AI workloads like LLM inference and training. And Apple Intelligence helps you write, express yourself, and get things done effortlessly with groundbreaking privacy protections at every step.*
- ALL-DAY BATTERY LIFE — MacBook Pro delivers the same exceptional performance whether it’s running on battery or plugged in.*
- MACOS RUNS APPS FAST — All your go-to apps run lightning fast in macOS, including built-in apps like FaceTime and Messages. Plus, built-in virus protection and free software updates help keep your Mac running smoothly and securely.
Checks to make in your own environment
- Confirm the TTL for every agent that holds approval state, and set it explicitly rather than relying on the default.
- Restart the container during a pending approval and confirm the resumed call still reaches the same task ID and reads the expected metadata.
- Keep request payloads and artifacts out of metadata and confirm they are retrievable from your own storage or checkpoint.
- Recheck the current Microsoft Learn pages for package versions and API signatures before you deploy, since the feature is in preview.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
