@speakai/mcp-server 2.0.0 → 2.0.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (3) hide show
  1. package/README.md +56 -17
  2. package/dist/index.js +87 -7
  3. package/package.json +1 -1
package/README.md CHANGED
@@ -39,11 +39,9 @@ The AI does the searching, summarizing, and citing. Your recordings stay in your
39
39
 
40
40
  ## Install (pick your tool)
41
41
 
42
- > **Two paths to install**: pick whichever feels easier. The one-click connect path approves access via a permission popup; the manual path pastes an API key into a header.
42
+ > **Start with the hosted connector.** Every assistant below connects to `https://api.speakai.co/v1/mcp` and signs you in through your browser on Speak AI's own page, where you can sign in or create a free account. There is no API key to copy. An API key setup is still available for local file upload, scripts and machines with no browser; see [Developer reference](#developer-reference-cli-api-advanced-setup).
43
43
 
44
- > **Don't know which one to pick?** Install it in the AI assistant you already use. Every option below connects to the same Speak AI workspace.
45
-
46
- > Speak AI's connector address (paste this into your AI tool's connector settings; it's the same idea as pasting a Zoom link into your calendar): `https://api.speakai.co/v1/mcp`
44
+ > **Don't know which one to pick?** Install it in the AI assistant you already use. Every option below connects to the same Speak AI account.
47
45
 
48
46
  ### Claude.ai (web)
49
47
 
@@ -128,32 +126,61 @@ Still stuck? Email [success@speakai.co](mailto:success@speakai.co).
128
126
 
129
127
  ### Claude Code (terminal)
130
128
 
131
- **Recommended: install from the official Claude Code plugin marketplace:**
129
+ Add the hosted connector, then sign in:
130
+
131
+ ```sh
132
+ claude mcp add --transport http speakai https://api.speakai.co/v1/mcp
133
+ claude mcp login speakai
134
+ ```
135
+
136
+ `claude mcp login` opens Speak AI's sign-in page in your browser. Sign in or create an account, then click **Allow**. You can also type `/mcp` inside Claude Code, pick **speakai** and choose to authenticate. On a machine with no browser (for example over SSH), run `claude mcp login speakai --no-browser`, open the printed link on another device, and paste the address your browser ends on back into the terminal.
137
+
138
+ Add `--scope user` to the first command to use Speak AI in every project instead of only the current one.
139
+
140
+ **Or install the plugin**, which adds the same hosted connector plus Speak AI skills:
132
141
 
133
142
  1. Add the official marketplace (one-time): `/plugin marketplace add claude-plugins-official`
134
143
  2. Install the plugin: `/plugin install speakai@claude-plugins-official`
135
144
  3. Activate it: `/reload-plugins`
136
- 4. Run the `getting-started` skill and paste your Speak AI API key. Generate one at [app.speakai.co/developers/apikeys](https://app.speakai.co/developers/apikeys).
145
+ 4. Type `/mcp`, pick **speakai** and sign in to Speak AI in your browser.
137
146
 
138
147
  If `/plugin install` doesn't find Speak AI, refresh the local catalog with `/plugin marketplace update claude-plugins-official` and retry.
139
148
 
140
149
  <details>
141
- <summary>Developer alternative: manual HTTP transport</summary>
150
+ <summary>Alternative: send an API key instead of signing in</summary>
151
+
152
+ ```sh
153
+ claude mcp add --transport http speakai https://api.speakai.co/v1/mcp \
154
+ --header "Authorization: Bearer $SPEAKAI_KEY"
155
+ ```
156
+
157
+ Set `SPEAKAI_KEY` in your shell first. Generate a key at [app.speakai.co/developers/apikeys](https://app.speakai.co/developers/apikeys).
142
158
 
143
- Skip the plugin and add the connector directly:
159
+ </details>
160
+
161
+ ### Codex (terminal)
144
162
 
145
163
  ```sh
146
- claude mcp add speakai --transport http --url https://api.speakai.co/v1/mcp
164
+ codex mcp add speakai --url https://api.speakai.co/v1/mcp
147
165
  ```
148
166
 
149
- Claude Code will open an OAuth window for sign-in. To bypass OAuth and pass a Bearer token instead:
167
+ Codex detects that Speak AI uses browser sign-in and opens Speak AI's sign-in page right away. Sign in or create an account, then click **Allow**. If the browser did not open, or you closed it before finishing, run:
150
168
 
151
169
  ```sh
152
- claude mcp add speakai --transport http --url https://api.speakai.co/v1/mcp \
153
- --header "Authorization: Bearer $SPEAKAI_KEY"
170
+ codex mcp login speakai
154
171
  ```
155
172
 
156
- Set `SPEAKAI_KEY` in your shell first, or paste your key inline. Generate a key at [app.speakai.co/developers/apikeys](https://app.speakai.co/developers/apikeys).
173
+ Check the connection with `codex mcp list`; Speak AI shows **OAuth** in the Auth column.
174
+
175
+ <details>
176
+ <summary>Alternative: send an API key instead of signing in</summary>
177
+
178
+ ```sh
179
+ export SPEAKAI_KEY="your-api-key"
180
+ codex mcp add speakai --url https://api.speakai.co/v1/mcp --bearer-token-env-var SPEAKAI_KEY
181
+ ```
182
+
183
+ Generate a key at [app.speakai.co/developers/apikeys](https://app.speakai.co/developers/apikeys).
157
184
 
158
185
  </details>
159
186
 
@@ -161,10 +188,22 @@ Set `SPEAKAI_KEY` in your shell first, or paste your key inline. Generate a key
161
188
 
162
189
  [![Add to Cursor](https://img.shields.io/badge/Cursor-Install_Speak_AI-000000?logo=cursor&logoColor=white&style=for-the-badge)](cursor://anysphere.cursor-deeplink/mcp/install?name=speakai&config=eyJ1cmwiOiJodHRwczovL2FwaS5zcGVha2FpLmNvL3YxL21jcCJ9)
163
190
 
164
- Click the button. Cursor registers itself automatically and opens the permission popup. Sign in to Speak AI and click **Allow**.
191
+ Click the button, or add this to `~/.cursor/mcp.json` (every project) or `.cursor/mcp.json` (one project):
192
+
193
+ ```json
194
+ {
195
+ "mcpServers": {
196
+ "speakai": {
197
+ "url": "https://api.speakai.co/v1/mcp"
198
+ }
199
+ }
200
+ }
201
+ ```
202
+
203
+ Then open **Cursor Settings → Tools & MCP**. Speak AI shows that it needs a login; click it to open Speak AI's sign-in page in your browser, sign in or create an account, and click **Allow**.
165
204
 
166
205
  <details>
167
- <summary>Developer alternative: manual setup with an API key</summary>
206
+ <summary>Alternative: local setup with an API key</summary>
168
207
 
169
208
  Use the manual stdio setup in the Developer reference at the bottom of this README.
170
209
 
@@ -275,11 +314,11 @@ Get a Speak AI API key at [app.speakai.co/developers/apikeys](https://app.speaka
275
314
 
276
315
  ### CLI / npm package
277
316
 
278
- The `@speakai/mcp-server` npm package provides:
317
+ The `@speakai/mcp-server` npm package runs on your own machine with a Speak AI API key. Use it when you need to upload local files from your computer, run scripts and pipelines, or connect a client that can't open a browser or doesn't support remote HTTP servers. It provides:
279
318
 
280
319
  - A CLI (`speakai-mcp`) for scripting and pipelines.
281
320
  - A stdio-mode MCP server for clients that don't support remote HTTP transport.
282
- - An auto-setup wizard that detects installed MCP clients and configures them.
321
+ - A setup wizard that shows the hosted connector command for the assistants it finds (Claude Code, Codex, Cursor), then offers to set up an API key for the local server and CLI.
283
322
 
284
323
  ```sh
285
324
  npm install -g @speakai/mcp-server
package/dist/index.js CHANGED
@@ -6369,10 +6369,10 @@ function register(server, client, options = {}) {
6369
6369
  if (modes.length !== 1) {
6370
6370
  return err(new Error("Send exactly one of imageUrl, imageBase64 or atSeconds."));
6371
6371
  }
6372
- const path4 = `/v1/media/${mediaId}/thumbnail`;
6372
+ const path5 = `/v1/media/${mediaId}/thumbnail`;
6373
6373
  try {
6374
6374
  if (imageUrl !== void 0) {
6375
- const result = await api.put(path4, { url: imageUrl });
6375
+ const result = await api.put(path5, { url: imageUrl });
6376
6376
  return ok(result.data);
6377
6377
  }
6378
6378
  if (imageBase64 !== void 0) {
@@ -6389,10 +6389,10 @@ function register(server, client, options = {}) {
6389
6389
  }
6390
6390
  const form = new FormData();
6391
6391
  form.append("image", new Blob([bytes], { type: mimeType }), `thumbnail.${THUMBNAIL_FILE_EXTENSIONS[mimeType]}`);
6392
- const result = await api.put(path4, form, { headers: { "Content-Type": "multipart/form-data" } });
6392
+ const result = await api.put(path5, form, { headers: { "Content-Type": "multipart/form-data" } });
6393
6393
  return ok(result.data);
6394
6394
  }
6395
- await api.put(path4, { seekSec: atSeconds });
6395
+ await api.put(path5, { seekSec: atSeconds });
6396
6396
  const deadline = Date.now() + THUMBNAIL_POLL_TIMEOUT_MS;
6397
6397
  while (Date.now() < deadline) {
6398
6398
  await new Promise((resolve) => setTimeout(resolve, THUMBNAIL_POLL_INTERVAL_MS));
@@ -11182,11 +11182,11 @@ function register16(server, client, options = {}) {
11182
11182
  }
11183
11183
  return step;
11184
11184
  };
11185
- const buildNodes = async (specs, path4) => {
11185
+ const buildNodes = async (specs, path5) => {
11186
11186
  const nodes2 = [];
11187
11187
  const folded = foldLegacyRunWhen(specs);
11188
11188
  for (const [index, spec] of folded.entries()) {
11189
- const where = `${path4}[${index}]`;
11189
+ const where = `${path5}[${index}]`;
11190
11190
  const stepId = nextStepId();
11191
11191
  const body2 = await buildStepBody(stepId, spec, where);
11192
11192
  if (String(spec.do ?? "") !== "branch") {
@@ -14923,6 +14923,73 @@ var init_format = __esm({
14923
14923
  }
14924
14924
  });
14925
14925
 
14926
+ // src/cli/hosted-setup.ts
14927
+ function onPath(command, envPath) {
14928
+ const exts = process.platform === "win32" ? [".cmd", ".exe", ""] : [""];
14929
+ return envPath.split(import_path2.default.delimiter).filter(Boolean).some((dir) => exts.some((ext) => (0, import_fs2.existsSync)(import_path2.default.join(dir, command + ext))));
14930
+ }
14931
+ function detectHostedClients(home = (0, import_os2.homedir)(), envPath = process.env.PATH ?? "") {
14932
+ const found = {
14933
+ "claude-code": onPath("claude", envPath) || (0, import_fs2.existsSync)(import_path2.default.join(home, ".claude")),
14934
+ codex: onPath("codex", envPath) || (0, import_fs2.existsSync)(import_path2.default.join(home, ".codex")),
14935
+ cursor: (0, import_fs2.existsSync)(import_path2.default.join(home, ".cursor"))
14936
+ };
14937
+ return ORDER.filter((c) => found[c]);
14938
+ }
14939
+ function hostedSetupText(detected) {
14940
+ const clients = detected.length > 0 ? detected : ORDER;
14941
+ const lines = [
14942
+ " Recommended: connect through the hosted connector.",
14943
+ " Your browser opens Speak AI's sign-in page, where you can sign in or create",
14944
+ " an account. No API key is needed.",
14945
+ ""
14946
+ ];
14947
+ if (detected.length === 0) {
14948
+ lines.push(" No AI assistant found on this machine. Pick yours:", "");
14949
+ }
14950
+ for (const c of clients) {
14951
+ lines.push(` ${HOSTED_SETUP[c].name}:`);
14952
+ for (const step of HOSTED_SETUP[c].steps) lines.push(` ${step}`);
14953
+ lines.push("");
14954
+ }
14955
+ lines.push(` Any other assistant: add ${HOSTED_MCP_URL} as a remote MCP server.`);
14956
+ return lines.join("\n");
14957
+ }
14958
+ var import_fs2, import_os2, import_path2, HOSTED_MCP_URL, HOSTED_SETUP, ORDER;
14959
+ var init_hosted_setup = __esm({
14960
+ "src/cli/hosted-setup.ts"() {
14961
+ "use strict";
14962
+ import_fs2 = require("fs");
14963
+ import_os2 = require("os");
14964
+ import_path2 = __toESM(require("path"));
14965
+ HOSTED_MCP_URL = "https://api.speakai.co/v1/mcp";
14966
+ HOSTED_SETUP = {
14967
+ "claude-code": {
14968
+ name: "Claude Code",
14969
+ steps: [
14970
+ `claude mcp add --transport http speakai ${HOSTED_MCP_URL}`,
14971
+ "claude mcp login speakai"
14972
+ ]
14973
+ },
14974
+ codex: {
14975
+ name: "Codex",
14976
+ steps: [
14977
+ `codex mcp add speakai --url ${HOSTED_MCP_URL}`,
14978
+ "codex mcp login speakai (only if the browser sign-in did not open or finish)"
14979
+ ]
14980
+ },
14981
+ cursor: {
14982
+ name: "Cursor",
14983
+ steps: [
14984
+ `Add to ~/.cursor/mcp.json: { "mcpServers": { "speakai": { "url": "${HOSTED_MCP_URL}" } } }`,
14985
+ "Open Cursor Settings > Tools & MCP and click the login prompt next to speakai"
14986
+ ]
14987
+ }
14988
+ };
14989
+ ORDER = ["claude-code", "codex", "cursor"];
14990
+ }
14991
+ });
14992
+
14926
14993
  // src/cli/index.ts
14927
14994
  var cli_exports = {};
14928
14995
  __export(cli_exports, {
@@ -15018,10 +15085,22 @@ function createCli() {
15018
15085
  saveConfig(cfg);
15019
15086
  printSuccess(`Base URL set to ${url}`);
15020
15087
  });
15021
- program.command("init").description("Interactive setup \u2014 configure API key and auto-detect MCP clients").action(async () => {
15088
+ program.command("init").description(
15089
+ "Interactive setup \u2014 shows the hosted connector command for your AI assistant, then offers API key setup"
15090
+ ).action(async () => {
15022
15091
  const rl = (0, import_readline.createInterface)({ input: process.stdin, output: process.stdout });
15023
15092
  const ask = (q) => new Promise((resolve) => rl.question(q, (a) => resolve(a.trim())));
15024
15093
  console.log("\n Speak AI MCP Server \u2014 Setup\n");
15094
+ console.log(hostedSetupText(detectHostedClients()));
15095
+ console.log(
15096
+ "\n An API key is only needed for the local server and the speakai-mcp CLI:\n uploading files from this computer, scripts, or a machine with no browser.\n"
15097
+ );
15098
+ const wantsKey = await ask(" Set up an API key as well? (y/N) ");
15099
+ if (wantsKey.toLowerCase() !== "y") {
15100
+ rl.close();
15101
+ printSuccess("Run the command above for your assistant to connect.");
15102
+ return;
15103
+ }
15025
15104
  const existingKey = resolveApiKey();
15026
15105
  let key = existingKey;
15027
15106
  if (existingKey) {
@@ -15934,6 +16013,7 @@ var init_cli = __esm({
15934
16013
  import_readline = require("readline");
15935
16014
  init_config();
15936
16015
  init_format();
16016
+ init_hosted_setup();
15937
16017
  init_media_utils();
15938
16018
  init_media3();
15939
16019
  }
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@speakai/mcp-server",
3
- "version": "2.0.0",
3
+ "version": "2.0.1",
4
4
  "mcpName": "io.github.speakai/mcp-server",
5
5
  "description": "Official Speak AI MCP Server — capture meetings, search thousands of recordings, run async voice and video surveys, create clips, and automate workflows from your AI assistant.",
6
6
  "homepage": "https://docs.speakai.co/mcp",