Every LLM API call resends your **entire conversation history**. A 20-turn chat costs 6x more per call than a 3-turn one, so you're paying for the
Promptthrift MCP MCP server is a locally run integration for AI assistants that speak the Model Context Protocol. Every LLM API call resends your entire conversation history. A 20-turn chat costs 6x more per call than a 3-turn one, so you're paying for the same old messages over and over.
Once Promptthrift MCP is connected, these are the calls the assistant has available:
Post — compression sanitizer** strips prompt injection patterns from summariesWebsite — https://aibranddefender.com/Guide — LINE AI Chatbot Guidepromptthrift_compress_history — Compress old turns into a smart summarypromptthrift_count_tokens — Track token usage & costs across 14 modelspromptthrift_suggest_model — Recommend cheapest model for the taskpromptthrift_pin_facts — Pin critical facts that survive compressionScenario — Without PromptThriftModel — Input $/MTokgemma-4-e2b — $0.00gemma-4-e4b — $0.00gemma-4-27b — $0.00You will need one environment variable: PROMPTTHRIFT_OLLAMA_URL. The server will not start without them, which is usually why the tools fail to appear on a first run. Keep credentials in your client's env block or a secrets manager rather than in a file you might commit.
Installation goes through your MCP client rather than a global install: point it at git on PyPI and it is fetched when the client starts. The copy-paste blocks for Claude Desktop, Claude Code and Cursor are further down this page.
This sits in the AI and media services group, where several servers overlap in what they claim to do but differ sharply once you actually set them up. Promptthrift MCP's toolset — Post, Website, Guide and 11 more — is a fair guide to whether it matches your workflow. It is maintained by woling-dev; worth a glance at recent repository activity before you build anything load-bearing on it.
This entry was verified against Promptthrift MCP's own documentation before publication; SyncDev keeps the directory reviewed rather than auto-generated.
| Tool | What it does |
|---|---|
| Post | compression sanitizer** strips prompt injection patterns from summaries |
| Website | https://aibranddefender.com/ |
| Guide | [LINE AI Chatbot Guide](https://aibranddefender.com/blog-line-ai-chatbot-guide.html) |
| promptthrift_compress_history | Compress old turns into a smart summary |
| promptthrift_count_tokens | Track token usage & costs across 14 models |
| promptthrift_suggest_model | Recommend cheapest model for the task |
| promptthrift_pin_facts | Pin critical facts that survive compression |
| Scenario | Without PromptThrift |
| Model | Input $/MTok |
| gemma-4-e2b | **$0.00** |
| gemma-4-e4b | **$0.00** |
| gemma-4-27b | **$0.00** |
| gemini-2.0-flash | $0.10 |
| gpt-4.1-nano | $0.10 |
{
"mcpServers": {
"promptthrift": {
"command": "uvx",
"args": ["git"],
"env": {
"PROMPTTHRIFT_OLLAMA_URL": "your-value"
}
}
}
}Add to claude_desktop_config.json, then restart Claude Desktop.
| Variable | Description | Required |
|---|---|---|
| PROMPTTHRIFT_OLLAMA_URL | Endpoint or connection string the server talks to. | Yes |
Build a programmable telecommunications stack for connecting telephony services with the Internet via a cloud-based utility.
Search built for AI, not humans — semantic web search that returns model-ready content, plus code context.
Answers, not links — delegate questions to Perplexity's search-grounded models and get cited responses back.
Give your assistant a voice — text-to-speech, voice cloning and audio tools from the ElevenLabs API.
Give your assistant a real code sandbox — isolated cloud VMs for actually running the code it writes.
The ML hub in your context window — search models, datasets, papers and run Spaces from the official server.