The two pieces already exist individually. There are plenty of options for both agent management (Conductor, Herdr, Omnigent) and microVM isolation (Docker Sandboxes, Firecracker, SmolVM). However, I did not find a project that combined them into one unified product.
Isolade gives each agent its own microVM. Microsandbox provides subsecond provisioning and domain-scoped secret substitution, so that the VM only sees placeholder values. Instead of a worktree, each agent gets a copy-on-write clone of the entire setup. This works well for setups involving multiple repositories, and cached dependencies mean agents can start working immediately without any per-worktree setup.
A single UI lets you work with multiple agents concurrently. You can mix Anthropic and OpenAI models and switch providers in the middle of a conversation. For example, I often have Sol review Opus's work. The system prompt is customizable, so for UI tasks, I ask agents to include actual screenshots of their changes or proposals in the conversation. This works much better than Claude's ASCII art.
You can use your existing subscriptions because Isolade runs the official Claude Code and Codex binaries.
Isolade is Apache-2.0 licensed and supports Apple Silicon macOS and Debian or Ubuntu with KVM. Looking forward to hearing your feedback."},"title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Isolade, a local-first coding agent workbench with secretless microVMs"},"url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/isolade/isolade"}},"_tags":["story","author_jachris","story_49168002","show_hn"],"author":"jachris","children":[49169561,49170604,49178635],"created_at":"2026-08-04T12:42:07Z","created_at_i":1785847327,"num_comments":4,"objectID":"49168002","points":3,"story_id":49168002,"story_text":"I built Isolade because I could not find an open-source, local-first workbench that runs coding agents in microVMs.
The two pieces already exist individually. There are plenty of options for both agent management (Conductor, Herdr, Omnigent) and microVM isolation (Docker Sandboxes, Firecracker, SmolVM). However, I did not find a project that combined them into one unified product.
Isolade gives each agent its own microVM. Microsandbox provides subsecond provisioning and domain-scoped secret substitution, so that the VM only sees placeholder values. Instead of a worktree, each agent gets a copy-on-write clone of the entire setup. This works well for setups involving multiple repositories, and cached dependencies mean agents can start working immediately without any per-worktree setup.
A single UI lets you work with multiple agents concurrently. You can mix Anthropic and OpenAI models and switch providers in the middle of a conversation. For example, I often have Sol review Opus's work. The system prompt is customizable, so for UI tasks, I ask agents to include actual screenshots of their changes or proposals in the conversation. This works much better than Claude's ASCII art.
You can use your existing subscriptions because Isolade runs the official Claude Code and Codex binaries.
Isolade is Apache-2.0 licensed and supports Apple Silicon macOS and Debian or Ubuntu with KVM. Looking forward to hearing your feedback.","title":"Show HN: Isolade, a local-first coding agent workbench with secretless microVMs","updated_at":"2026-08-05T04:41:02Z","url":"https://github.com/isolade/isolade"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"joostdevries"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"Interesting!\nPersonally I'm not eager to spend tokens on having my own harness, sandbox etc.\nIt's much much more feasible by spending tokens. But it does feel like a distraction and that I end up owning it. As in: a continuing distraction.
So I've been experimenting for these purposes with omnigent as a way to be less locked into a single provider and for its sandbox abstraction. And I've also tried openshell for sandboxing. Hoping those two will keep improving."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Pi pod \u2013 Run your pi coding agent in sandboxes on your own server"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://pipod.dev/"}},"_tags":["comment","author_joostdevries","story_49937304"],"author":"joostdevries","comment_text":"Interesting!\nPersonally I'm not eager to spend tokens on having my own harness, sandbox etc.\nIt's much much more feasible by spending tokens. But it does feel like a distraction and that I end up owning it. As in: a continuing distraction.
So I've been experimenting for these purposes with omnigent as a way to be less locked into a single provider and for its sandbox abstraction. And I've also tried openshell for sandboxing. Hoping those two will keep improving.","created_at":"2026-10-04T07:25:35Z","created_at_i":1791098735,"objectID":"49951491","parent_id":49937304,"story_id":49937304,"story_title":"Show HN: Pi pod \u2013 Run your pi coding agent in sandboxes on your own server","story_url":"https://pipod.dev/","updated_at":"2026-10-04T11:48:56Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"mbil"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"Also maybe https://omnigent.ai/"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Offrun \u2013 manage every coding agent from one workspace"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://offrun.dev/"}},"_tags":["comment","author_mbil","story_49942434"],"author":"mbil","children":[49946332],"comment_text":"Also maybe https://omnigent.ai/","created_at":"2026-10-03T17:32:14Z","created_at_i":1791048734,"objectID":"49946125","parent_id":49945907,"story_id":49942434,"story_title":"Show HN: Offrun \u2013 manage every coding agent from one workspace","story_url":"https://offrun.dev/","updated_at":"2026-10-03T17:59:07Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jeffnash"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"I agree the gap is narrowing quickly, especially with workflows in CC. For me, a few large advantages still remain:
1/ Allowing me to easily plug in any harness, using any provider, and make it a first-class worker. Omnigent has out of the box ACP support and it's trivial to use that to add first-class support for any harness out there. I love the ability to have CC + Opus plan, Codex + Luna implement, Pi + Qwen 3.8 give a tie-breaking opinion on a design decision that Opus flagged and Grok and Codex couldn't agree on, all orchestrated by a model of my choosing from any provider using Omnigent's main agent harness.
2/ Reusable agent systems rather than just reusable workflows. You can define agents in YAML whose subagents embody particular roles, with different models/harnesses, skills, plugins, tools, etc. preconfigured for each one.
Of course, claude workflows are now durable but Omnigent's agnt definitions are a bit more abstract in that they define the subagents that are available and how they should work by default rather than the workflow itself (i.e. the specific JTBD). If I have a common workflow that consists of, for example, Sol + Codex writing some script to scrape some data, Pi + a cheap DeepSeek-tier model formatting that data en masse, then Fable + CC doing some advanced analysis on it, I can embody that with a yaml agent definition that I can then use to run with my task of the day as a prompt. All of this is orchestrated by a model of my choice using Omnigent's harness.
This might look like: 'smart scraping agent with all sorts of scraping skills and tools pre-loaded', a 'bulk data processing agent with a cheap, fast model and plenty of pandas/numpy skills preloaded', and 'frontier model to interpret and reason on the implications of the processed data'. The main agent would have instructions about the general workflow of such tasks and when to invoke and delegate tasks to which subagent. The definition describes the workers available to the orchestrator and how they should generally behave, rather than hard-coding the workflow itself. I love that I can create those definitions and re-use them.
All that said, I am sure the labs will come up with their own similar products to (2) (e.g. dots today). I also recently noticed that Claude Code now has subagent 'teams' rather than just 'general-purpose'/'explore' subagents, and these seem to be longer-lived. This seems to be encroaching on the agent yaml definitions, albeit with less fine-grained control on my end. Therefore, the tl;dr (for me at least) is vendor neutrality; I don't think we'll ever see a product coming out of a frontier lab that eagerly delegates a task to their competitor's model (and bank account)."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Raven \u2013 The harness of harnesses, built for RSI"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/EverMind-AI/Raven"}},"_tags":["comment","author_jeffnash","story_49890647"],"author":"jeffnash","comment_text":"I agree the gap is narrowing quickly, especially with workflows in CC. For me, a few large advantages still remain:
1/ Allowing me to easily plug in any harness, using any provider, and make it a first-class worker. Omnigent has out of the box ACP support and it's trivial to use that to add first-class support for any harness out there. I love the ability to have CC + Opus plan, Codex + Luna implement, Pi + Qwen 3.8 give a tie-breaking opinion on a design decision that Opus flagged and Grok and Codex couldn't agree on, all orchestrated by a model of my choosing from any provider using Omnigent's main agent harness.
2/ Reusable agent systems rather than just reusable workflows. You can define agents in YAML whose subagents embody particular roles, with different models/harnesses, skills, plugins, tools, etc. preconfigured for each one.
Of course, claude workflows are now durable but Omnigent's agnt definitions are a bit more abstract in that they define the subagents that are available and how they should work by default rather than the workflow itself (i.e. the specific JTBD). If I have a common workflow that consists of, for example, Sol + Codex writing some script to scrape some data, Pi + a cheap DeepSeek-tier model formatting that data en masse, then Fable + CC doing some advanced analysis on it, I can embody that with a yaml agent definition that I can then use to run with my task of the day as a prompt. All of this is orchestrated by a model of my choice using Omnigent's harness.
This might look like: 'smart scraping agent with all sorts of scraping skills and tools pre-loaded', a 'bulk data processing agent with a cheap, fast model and plenty of pandas/numpy skills preloaded', and 'frontier model to interpret and reason on the implications of the processed data'. The main agent would have instructions about the general workflow of such tasks and when to invoke and delegate tasks to which subagent. The definition describes the workers available to the orchestrator and how they should generally behave, rather than hard-coding the workflow itself. I love that I can create those definitions and re-use them.
All that said, I am sure the labs will come up with their own similar products to (2) (e.g. dots today). I also recently noticed that Claude Code now has subagent 'teams' rather than just 'general-purpose'/'explore' subagents, and these seem to be longer-lived. This seems to be encroaching on the agent yaml definitions, albeit with less fine-grained control on my end. Therefore, the tl;dr (for me at least) is vendor neutrality; I don't think we'll ever see a product coming out of a frontier lab that eagerly delegates a task to their competitor's model (and bank account).","created_at":"2026-09-29T19:03:36Z","created_at_i":1790708616,"objectID":"49898702","parent_id":49893841,"story_id":49890647,"story_title":"Show HN: Raven \u2013 The harness of harnesses, built for RSI","story_url":"https://github.com/EverMind-AI/Raven","updated_at":"2026-09-29T19:12:39Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"marginalx"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"I'm curious if you have a few mins for feedback, what are the top 2 things here that omnigent does that is significantly better for you than latest cc/codex which can launch subagents, auto save memory of a project.
I'm wondering that as these core tools continue to enhance and add these capabilities, how much of a benefit these meta harnesses actually provide."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Raven \u2013 The harness of harnesses, built for RSI"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/EverMind-AI/Raven"}},"_tags":["comment","author_marginalx","story_49890647"],"author":"marginalx","children":[49898702],"comment_text":"I'm curious if you have a few mins for feedback, what are the top 2 things here that omnigent does that is significantly better for you than latest cc/codex which can launch subagents, auto save memory of a project.
I'm wondering that as these core tools continue to enhance and add these capabilities, how much of a benefit these meta harnesses actually provide.","created_at":"2026-09-29T14:23:11Z","created_at_i":1790691791,"objectID":"49893841","parent_id":49892818,"story_id":49890647,"story_title":"Show HN: Raven \u2013 The harness of harnesses, built for RSI","story_url":"https://github.com/EverMind-AI/Raven","updated_at":"2026-09-29T19:04:25Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jeffnash"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"Reminds me a lot of omnigent (which I am a huge fan of) with a persistent memory layer. Unlike omnigent's subagent threads, the DAG it uses to coordinate other harnesses doesn't look to be durable; I am curious as to whether this is by design or is a forthcoming feature, as this essentially makes or breaks my use case of long-running project-sized implementation sessions.
In any event, it's great to see competition in this meta-harness space, which is likely one that none of the frontier labs will touch since it, by definition, would utilize their competitors' products."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Show HN: Raven \u2013 The harness of harnesses, built for RSI"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/EverMind-AI/Raven"}},"_tags":["comment","author_jeffnash","story_49890647"],"author":"jeffnash","children":[49893841,49893848,49902198],"comment_text":"Reminds me a lot of omnigent (which I am a huge fan of) with a persistent memory layer. Unlike omnigent's subagent threads, the DAG it uses to coordinate other harnesses doesn't look to be durable; I am curious as to whether this is by design or is a forthcoming feature, as this essentially makes or breaks my use case of long-running project-sized implementation sessions.
In any event, it's great to see competition in this meta-harness space, which is likely one that none of the frontier labs will touch since it, by definition, would utilize their competitors' products.","created_at":"2026-09-29T13:22:09Z","created_at_i":1790688129,"objectID":"49892818","parent_id":49890647,"story_id":49890647,"story_title":"Show HN: Raven \u2013 The harness of harnesses, built for RSI","story_url":"https://github.com/EverMind-AI/Raven","updated_at":"2026-09-30T08:31:56Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"chachachainsaw"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"I use omnigent with claude code and codex. I also built a tool (https://frontmatter.news) to help me figure out what similarities and differences there are between tools so my agents behave the same way across tools."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Ask HN: What's your AI coding set up?"}},"_tags":["comment","author_chachachainsaw","story_49636239"],"author":"chachachainsaw","comment_text":"I use omnigent with claude code and codex. I also built a tool (https://frontmatter.news) to help me figure out what similarities and differences there are between tools so my agents behave the same way across tools.","created_at":"2026-09-18T13:25:10Z","created_at_i":1789737910,"objectID":"49754078","parent_id":49636239,"story_id":49636239,"story_title":"Ask HN: What's your AI coding set up?","updated_at":"2026-09-18T13:29:05Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"vehemenz"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"Given that like omnigent and qm already integrate with Slack, it's not a huge surprise that Slack is building agent capabilities directly in.
Personally, I'm getting ecosystem fatigue. Who even has the time to set up pilots to feature-test all these different systems?"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Slack Code"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.salesforce.com/introducing-slack-code/?bc=HL"}},"_tags":["comment","author_vehemenz","story_49374965"],"author":"vehemenz","comment_text":"Given that like omnigent and qm already integrate with Slack, it's not a huge surprise that Slack is building agent capabilities directly in.
Personally, I'm getting ecosystem fatigue. Who even has the time to set up pilots to feature-test all these different systems?","created_at":"2026-08-20T15:09:18Z","created_at_i":1787238558,"objectID":"49375743","parent_id":49374965,"story_id":49374965,"story_title":"Slack Code","story_url":"https://www.salesforce.com/introducing-slack-code/?bc=HL","updated_at":"2026-08-21T05:20:34Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"Cameri"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"How is OneCLI different from Databrick's Omnigent?"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Launch HN: OneCLI (YC S26) \u2013 OSS sandboxed agent harness for teams"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://github.com/onecli/onecli"}},"_tags":["comment","author_Cameri","story_49363710"],"author":"Cameri","comment_text":"How is OneCLI different from Databrick's Omnigent?","created_at":"2026-08-20T00:41:51Z","created_at_i":1787186511,"objectID":"49369075","parent_id":49363710,"story_id":49363710,"story_title":"Launch HN: OneCLI (YC S26) \u2013 OSS sandboxed agent harness for teams","story_url":"https://github.com/onecli/onecli","updated_at":"2026-08-20T00:45:26Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"vira28"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"Reads like an add to Omnigent or whatever harness (wait it\u2019s meta harness?."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Managing AI Coding Costs at Scale"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.databricks.com/blog/managing-ai-coding-costs-scale"}},"_tags":["comment","author_vira28","story_49214468"],"author":"vira28","comment_text":"Reads like an add to Omnigent or whatever harness (wait it\u2019s meta harness?.","created_at":"2026-08-08T01:30:00Z","created_at_i":1786152600,"objectID":"49218087","parent_id":49214468,"story_id":49214468,"story_title":"Managing AI Coding Costs at Scale","story_url":"https://www.databricks.com/blog/managing-ai-coding-costs-scale","updated_at":"2026-08-08T06:29:44Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"jvican"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"Omnigent seems to compete more against Orca https://github.com/stablyai/orca\nThey both went to be the Agent IDE layer, where you come with your tasks and everything is taken care of. I've been using Orca for a handful of tasks and have been largely enjoying it. My default barebones workflow is ghostty + zmx on ssh connections."},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Managing AI Coding Costs at Scale"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.databricks.com/blog/managing-ai-coding-costs-scale"}},"_tags":["comment","author_jvican","story_49214468"],"author":"jvican","children":[49216975],"comment_text":"Omnigent seems to compete more against Orca https://github.com/stablyai/orca\nThey both went to be the Agent IDE layer, where you come with your tasks and everything is taken care of. I've been using Orca for a handful of tasks and have been largely enjoying it. My default barebones workflow is ghostty + zmx on ssh connections.","created_at":"2026-08-07T21:33:00Z","created_at_i":1786138380,"objectID":"49216491","parent_id":49216028,"story_id":49214468,"story_title":"Managing AI Coding Costs at Scale","story_url":"https://www.databricks.com/blog/managing-ai-coding-costs-scale","updated_at":"2026-08-08T02:25:44Z"},{"_highlightResult":{"author":{"matchLevel":"none","matchedWords":[],"value":"ankitmathur"},"comment_text":{"fullyHighlighted":false,"matchLevel":"full","matchedWords":["omnigent"],"value":"Omnigent and OpenRouter are different in the sense that OpenRouter is where you can go to call the actual model but Omnigent is intended to be the place where you go describe the high level task to be done, and work is farmed out to various harnesses and models. Those sandboxes can themselves be using OpenRouter for capacity!
We're calling the layer coordinating harnesses "meta-harness'"},"story_title":{"matchLevel":"none","matchedWords":[],"value":"Managing AI Coding Costs at Scale"},"story_url":{"matchLevel":"none","matchedWords":[],"value":"https://www.databricks.com/blog/managing-ai-coding-costs-scale"}},"_tags":["comment","author_ankitmathur","story_49214468"],"author":"ankitmathur","children":[49216491],"comment_text":"Omnigent and OpenRouter are different in the sense that OpenRouter is where you can go to call the actual model but Omnigent is intended to be the place where you go describe the high level task to be done, and work is farmed out to various harnesses and models. Those sandboxes can themselves be using OpenRouter for capacity!
We're calling the layer coordinating harnesses "meta-harness'","created_at":"2026-08-07T20:46:31Z","created_at_i":1786135591,"objectID":"49216028","parent_id":49215966,"story_id":49214468,"story_title":"Managing AI Coding Costs at Scale","story_url":"https://www.databricks.com/blog/managing-ai-coding-costs-scale","updated_at":"2026-08-08T17:24:31Z"}],"hitsPerPage":20,"nbHits":752,"nbPages":38,"page":0,"params":"query=omnigent&advancedSyntax=true&analyticsTags=backend","processingTimeMS":7,"processingTimingsMS":{"_request":{"queue":2,"roundTrip":15},"fetch":{"query":5,"total":6},"total":7},"query":"omnigent","serverTimeMS":10}