Wednesday at 02:32 AM4 days No it is a mix - usually the first few in a reply is program name with link, then some program numbers with link and in the end it is text only - it is like it runs out of links :-) Edited Wednesday at 02:34 AM4 days by Panda88
Wednesday at 03:31 PM4 days Author Hello everyone, please upgrade to 2.4.0 ... major improvements in cache management and token usage.With kind regards,Michel
Wednesday at 03:44 PM4 days 12 minutes ago, Michel Kohanim said:Hello everyone, please upgrade to 2.4.0 ... major improvements in cache management and token usage.With kind regards,MichelDone. Took about 10 seconds.
Wednesday at 08:27 PM4 days On 11/12/2026 at 5:13 PM, Michel Kohanim said:Hello everyone,Some exciting stuff:User guide! [ link ]Migration of Python from 311 to 312Added a tool that manages 90% of github setup so that all you need to do is to a) only once: setup your github account b) create a repo and set it up with the created by eisy-ai, and c) point eisy-ai to that repo. Then just to git stuff directly from eisy-ai such as add, commit, push, fetch ...IMPORTANT WARNING - you need the latest udx and eisy-ui. Upgrading from Staging will upgrade your OS to 14.5 AND your python to 312. Please be patient. It may take up to 2 hours.With kind regards,MichelOk I finally got around to doing this today. I have to say, for such a major update, it went really well! As expected, I had to wait a bit more while it reinstalled plugins, and needed to restart some, but in the end, everything was working. I then proceeded to migrate the plugins away from PG3x, one by one. Again, smooth sailing. There is only one minor issue remaining: It won't migrate the Virtual plugin because it wasn't installed from the production store. Indeed it was installed from the beta store. Thing is, the beta version is 3.2.3 and production store version is 3.2.2, so even though I go to the production store and click on "update" in the plugin's configuration screen (and it appears to be udating), the version remains the same, probably because it sees that the production version number is lower. How do I get around that?
Wednesday at 08:29 PM4 days Author @Guy Lavoie thank you for the report!Regarding the issue, please submit a ticket.With kind regards,Michel
Wednesday at 08:40 PM4 days 10 minutes ago, Michel Kohanim said:@Guy Lavoie thank you for the report!Regarding the issue, please submit a ticket.With kind regards,MichelDone.
Wednesday at 09:01 PM4 days Author Hello everyone,Initial analysis of the new version showed 70% improvement in token/context usage. Added a little more housekeeping, and removes things that made processing more difficult (max_turns), and in one case we got 96% cache hit developing a plugin for Rheem Econet! So, if you have not done so yet, please upgrade to the latest to save on tokens/costs.With kind regards,Michel
Wednesday at 09:18 PM4 days Done. Can I assume that if I have 2 different easy's at 2 different locations, that each eisy ai plugin requires a different api for the configuration?
Wednesday at 09:18 PM4 days 12 minutes ago, Michel Kohanim said:Hello everyone,Initial analysis of the new version showed 70% improvement in token/context usage. Added a little more housekeeping, and removes things that made processing more difficult (max_turns), and in one case we got 96% cache hit developing a plugin for Rheem Econet!So, if you have not done so yet, please upgrade to the latest to save on tokens/costs.With kind regards,Michel@Michel Kohanim Glad to hear this!Do we need to make any adjustments to the Configuration Json? A quick check shows that there is a setting "max_turns": 20?Also, what is the current recommendation for history_token_budget? From reading your posts, it looks like <50000 - 20000?
Wednesday at 09:28 PM4 days Author @DennisC ,Good questions:max_turns is no longer used. It does not hurt if you keep it. My suggestions:Customer mode:max_tokens: 20000history_token_budget: 60000Developer mode:max_tokens: 64000history_token_budget: 100000You can always adjust history_token_budgets if you your conversations are compacted too frequently.With kind regards,Michel
Thursday at 04:08 PM3 days 18 hours ago, Michel Kohanim said:@DennisC ,Good questions:max_turns is no longer used. It does not hurt if you keep it. My suggestions:Customer mode:max_tokens: 20000history_token_budget: 60000Developer mode:max_tokens: 64000history_token_budget: 100000You can always adjust history_token_budgets if you your conversations are compacted too frequently.With kind regards,MichelI can confirm, these settings drastically reduced cost!
Thursday at 04:12 PM3 days Author @DennisC ,Thank you so much for the feedback. A minor issue is that the compaction doesn't summarize tool calls, so you will be seeing something like this after compaction:I told you x, but I was lying ... do you want to verify?Another minor issue is that the compaction happens way too frequently (for me). We are working on both of these issues while keeping the costs down. I shall keep you posted.With kind regards,Michel
Thursday at 04:21 PM3 days 5 minutes ago, Michel Kohanim said:@DennisC ,Thank you so much for the feedback. A minor issue is that the compaction doesn't summarize tool calls, so you will be seeing something like this after compaction:I told you x, but I was lying ... do you want to verify?Another minor issue is that the compaction happens way too frequently (for me).We are working on both of these issues while keeping the costs down. I shall keep you posted.With kind regards,MichelI have been working and I have exchanged over a dozen prompts so far. I am currently only at 60% for compaction. In my experience so far, this is unheard of! Compaction was occurring every two - three prompts. I was just going to compact, but maybe I will hold off.........lol
Thursday at 05:09 PM3 days @Michel Kohanim For what it's worth, I compacted 3 times with no funny business.
Thursday at 05:11 PM3 days Author Well, you need to first make it fail by going back and forth between plugins in the same session!With kind regards,Michel
Thursday at 07:41 PM3 days Yesterday I tried yet a different approach to creating a plugin, a bolder one. I spelled out everything I wanted in a single long request. It was about 7 lines long in the chat window. I told it the name I wanted, how many nodes to create, characteristics (raw number, increment once per second if non-zero), control buttons (stop, pause, etc) and pressed enter. It worked from the first try. So don't be afraid to be fussy and detailed!
Thursday at 08:18 PM3 days Author While testing long context, multi-turn, token-gobbling scenarios today, I had it create a chess plugin! And it won!With kind regards,Michel
Thursday at 11:31 PM2 days 2 hours ago, photogeek54 said:Look out humanity the AIs are winning 😝Revenge of the Machines!
Thursday at 11:51 PM2 days 3 hours ago, Michel Kohanim said:While testing long context, multi-turn, token-gobbling scenarios today, I had it create a chess plugin! And it won!With kind regards,MichelDo you enter moves with a keypadlinc? Lol
Friday at 02:47 AM2 days Author 2 hours ago, Guy Lavoie said:Do you enter moves with a keypadlinc? LolNo, there's an API for moves. Next is to have it make a UI ... With kind regards,Michel
Friday at 04:34 AM2 days Author Hello everyone,Version 2.4.2 is released and it's a major improvement for token management for Claude with more than 90% cache hit rate in majority of cases. If you wish, you can also increase token_history_budget to 250K. The only charge will be after 1+ hour of inactivity for a LONG session. Otherwise, it should be smooth sailing. Most interesting thing is /session-stats that gives you a breakdown of everything you need to know regarding your tokens.Here are the release notes:- Compact a Claude session by forking the live conversation. The summarizing request now reuses the turn's own system prompt, history and tools (so the already-cached history is read at the cache-read price) with a final instruction to summarize, and the model sees every tool result; on success the entire history is replaced by that one summary. Any other provider, or a fork that fails, keeps the standalone summarizer (oldest turns summarized, newest half-budget kept), which now also sees one line per tool call with a short result excerpt. `/compact-session` passes the connection context through so it can fork too. - Fix a cache regression on every idle gap: Claude cache breakpoints now all use a fixed 1-hour TTL (tools, every system block, and the conversation, which was hardcoded to 5 minutes), so a pause of up to an hour between turns reads the cache instead of rewriting the whole history. **Upgrade note:** the `cache_ttl` profile option is removed; a leftover `"cache_ttl"` (the example configs set `"5m"`) is silently ignored, so those profiles now cache for 1 hour, where writes cost 2x instead of 1.25x (reads are unchanged at 0.1x). - Remove the installed-plugins list from the plugin-authoring system prompt. It was the last block before the conversation, so any plugin install/start/stop rewrote the whole cached conversation (about 25K tokens in one logged session). The model calls `list_installed_plugins` for live state instead. - Add `/session-stats`: shows what the next request is made of (each system section, the tools, and history split into compacted summary, conversation text and tool calls/results, scaled to the last real request's input total when comparable), the last request's cache reads/writes and hit rate, session totals with an estimated cost for known models, compactions (fork/summary/truncated, before and after), and the last five requests. It makes no model call. Per-request token usage is now recorded for every provider and saved with the session instead of only appearing in the prompt log. New optional profile setting `context_window` (shown as the share of the window used) for models the built-in table doesn't know. - Log (without compacting) when a single turn grows past twice `history_token_budget`: a `would compact mid-turn` warning, a `mid_turn_over_budget` prompt-log line and a counter in `/session-stats`. Compaction is still only evaluated at the start of a turn. - Scrub API keys and tokens out of every log record, including `debug.log`: udi_interface's own handler logged each Polyglot message with the plugin's custom parameters, provider API keys included, which the existing redaction never reached. A logging record factory now removes keys by shape (`sk-ant-`, `sk-`, `xai-`, `gsk_`, `AIza`, `Bearer`) and quoted values after fields such as `api_key`/`client_secret`/`password`. Session persistence is unaffected. Lines written by earlier versions stay in old logs, so rotate any key that appeared in one. - Add the `mdns_search` plugin-authoring tool (a stdlib-only mDNS browser that queries from an ephemeral port so it never competes with the eisy's own responder) and mDNS/Bonjour support in generated plugins (`server_entry.mdns_enabled` wires a handler forwarding to the required `handle_mdns_results` override). - Put a blank line after the history-budget and compaction notices so the reply no longer runs onto the notice's last word.With kind regards,Michel
Friday at 08:49 PM2 days Made another personal advancement today: setting up github and pushed my three plugins. Thanks in large part to @apostolakisl and his excellent writeup on setting up github (see page 6 of this thread). His instructions from step 3 onwards have now mostly been replaced with a streamlined process that was added to eisy-ai. Just give it the url and deploy key when asked and it gets done. 🙂The next step is to make these plugins available to community, Not sure how we proceed to submitting that and getting vetted, but I'm guessing it's by opening a ticket, so I'll go do that now. Thanks for the help in getting me there!
Friday at 08:51 PM2 days Author @Guy Lavoie ,Thank you! Please submit a ticket and we'll walk you through it next week.With kind regards,Michel
Friday at 08:57 PM2 days Author @DennisC Not entirely true. You cannot publish if you're not setup as a developer. So, a) verify identity b) set you up as a developer. So, right now, non-developer customers do NOT get the Publish button.With kind regards,Michel
Create an account or sign in to comment