deepseek-ai/awesome-deepseek-integration
1
0
Fork 0
mirror of https://github.com/deepseek-ai/awesome-deepseek-integration.git synced 2026-11-05 00:57:30 +00:00
Code Issues 184 Projects Packages Wiki Activity

[PR #751] Add AiRC Orchestration: multi-agent workflows with DeepSeek #728

New issue
Open
opened 2026-10-26 10:01:00 +00:00 by bsakaguchi · 0 comments
bsakaguchi commented 2026-10-26 10:01:00 +00:00
Owner
Copy link

📋 Pull Request Information

Original PR: https://github.com/deepseek-ai/awesome-deepseek-integration/pull/751
Author: @AiRC-ai
Created: 10/2/2026
Status: 🔄 Open

Base: main ← Head: codex/add-airc-orchestration

📝 Commits (2)

  • cda04d1 docs: add AiRC Orchestration to Applications
  • 598fd51 docs: explain AiRC cross-provider agent orchestration

📊 Changes

5 files changed (+25 additions, -0 deletions)

View changed files

📝 README.md (+5 -0)
📝 README_cn.md (+5 -0)
📝 README_es.md (+5 -0)
📝 README_ja.md (+5 -0)
📝 README_zh_tw.md (+5 -0)

📄 Description

Adds AiRC Orchestration to Applications in all five language READMEs, with localized descriptions of its agent-orchestration capabilities.

AiRC lets DeepSeek act as the primary coordinating agent or as a worker in a configurable team for coding, research and automation. The direct DeepSeek integration uses the OpenAI-compatible API at https://api.deepseek.com, the user's API key, and model discovery. Teams can use several DeepSeek workers or mix DeepSeek with other supported local and hosted models.

What the orchestration adds

  • Configurable agent teams: choose a provider, model and reasoning setting for each worker slot, or inherit the primary task's model. Multiple slots can use the same model, and teams can mix supported providers.
  • Parallel tool-using workers: the primary agent assigns named, scoped tasks. Each worker runs its own agent loop and conversation with configured tools and a working directory. The primary can continue independent work, steer or resume workers, and collect their results.
  • A persistent project workflow: goals, plans, task history and worker activity stay connected to the project. The desktop shows worker identities, tool activity and status.
  • Optional supervision: a separately selected reviewer model checks recorded work against the goal and plan and can provide bounded corrective guidance.
  • Tools beyond chat: MCP extensions, file and shell workflows, a task-scoped browser, and recurring automations support coding, research and operational tasks. An optional Vision Bridge lets a configured vision model describe images for a text-only model.

For example, a primary agent can split a repository investigation among architecture, test and documentation workers, then reconcile their findings into an implementation plan. This illustrates the intended workflow; it is not a new performance benchmark.

Parallelism is configurable, not unlimited. AiRC supports a global sub-agent concurrency setting up to 32; per-task/background limits, provider stream budgets and actual host/account capacity can lower simultaneous execution. Tool use and task success depend on the selected models, configured capabilities and permissions.

Availability and validation

Public downloads are available for macOS Apple silicon, Windows x64 and Debian-family Linux amd64. AiRC is proprietary; its public license permits personal evaluation and testing. Provider charges depend on the chosen service.

The orchestration claims were checked against the source revision identified by public release 1.45.0+airc115, including worker creation, model routing, background execution, persistent child state, supervision and concurrency enforcement. Existing regression tests were inspected, not run. No new runtime benchmark or application build is claimed. Directory formatting and public links were checked; the submitted change is documentation only.

Submitted by AiRC's developer account with AI assistance. No upstream endorsement is implied.

🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.

## 📋 Pull Request Information **Original PR:** https://github.com/deepseek-ai/awesome-deepseek-integration/pull/751 **Author:** [@AiRC-ai](https://github.com/AiRC-ai) **Created:** 10/2/2026 **Status:** 🔄 Open **Base:** `main` ← **Head:** `codex/add-airc-orchestration` --- ### 📝 Commits (2) - [`cda04d1`](https://github.com/deepseek-ai/awesome-deepseek-integration/commit/cda04d1d682df01a451deeed02fea0fd6d68e997) docs: add AiRC Orchestration to Applications - [`598fd51`](https://github.com/deepseek-ai/awesome-deepseek-integration/commit/598fd51702e56e796fae739927b8435ad291bbdc) docs: explain AiRC cross-provider agent orchestration ### 📊 Changes **5 files changed** (+25 additions, -0 deletions) <details> <summary>View changed files</summary> 📝 `README.md` (+5 -0) 📝 `README_cn.md` (+5 -0) 📝 `README_es.md` (+5 -0) 📝 `README_ja.md` (+5 -0) 📝 `README_zh_tw.md` (+5 -0) </details> ### 📄 Description Adds [AiRC Orchestration](https://github.com/AiRC-ai/AiRC-Orchestration) to Applications in all five language READMEs, with localized descriptions of its agent-orchestration capabilities. AiRC lets DeepSeek act as the primary coordinating agent or as a worker in a configurable team for coding, research and automation. The direct DeepSeek integration uses the OpenAI-compatible API at `https://api.deepseek.com`, the user's API key, and model discovery. Teams can use several DeepSeek workers or mix DeepSeek with other supported local and hosted models. ## What the orchestration adds - **Configurable agent teams:** choose a provider, model and reasoning setting for each worker slot, or inherit the primary task's model. Multiple slots can use the same model, and teams can mix supported providers. - **Parallel tool-using workers:** the primary agent assigns named, scoped tasks. Each worker runs its own agent loop and conversation with configured tools and a working directory. The primary can continue independent work, steer or resume workers, and collect their results. - **A persistent project workflow:** goals, plans, task history and worker activity stay connected to the project. The desktop shows worker identities, tool activity and status. - **Optional supervision:** a separately selected reviewer model checks recorded work against the goal and plan and can provide bounded corrective guidance. - **Tools beyond chat:** MCP extensions, file and shell workflows, a task-scoped browser, and recurring automations support coding, research and operational tasks. An optional Vision Bridge lets a configured vision model describe images for a text-only model. For example, a primary agent can split a repository investigation among architecture, test and documentation workers, then reconcile their findings into an implementation plan. This illustrates the intended workflow; it is not a new performance benchmark. Parallelism is configurable, not unlimited. AiRC supports a global sub-agent concurrency setting up to 32; per-task/background limits, provider stream budgets and actual host/account capacity can lower simultaneous execution. Tool use and task success depend on the selected models, configured capabilities and permissions. ## Availability and validation [Public downloads](https://github.com/AiRC-ai/AiRC-Orchestration/releases/latest) are available for macOS Apple silicon, Windows x64 and Debian-family Linux amd64. AiRC is proprietary; its [public license](https://github.com/AiRC-ai/AiRC-Orchestration/blob/main/LICENSE) permits personal evaluation and testing. Provider charges depend on the chosen service. The orchestration claims were checked against the source revision identified by public release `1.45.0+airc115`, including worker creation, model routing, background execution, persistent child state, supervision and concurrency enforcement. Existing regression tests were inspected, not run. No new runtime benchmark or application build is claimed. Directory formatting and public links were checked; the submitted change is documentation only. Submitted by AiRC's developer account with AI assistance. No upstream endorsement is implied. --- <sub>🔄 This issue represents a GitHub Pull Request. It cannot be merged through Gitea due to API limitations.</sub>
Sign in to join this conversation.
No Branch/Tag specified
Branches Tags
main
No results found.
Labels
Clear labels   
pull-request
Mirrored from GitHub Pull Request

No labels
pull-request
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
deepseek-ai/awesome-deepseek-integration#728
No description provided.